Artificial Intelligence in Genomics Market worth $1,671 million by 2025 says MarketsandMarkets – Benzinga

Chicago, Oct. 06, 2022 (GLOBE NEWSWIRE) -- According to the new market research report by MarketsandMarkets, theArtificial Intelligence In Genomics Market is projected to reach USD 1,671 million by 2025 from USD 202 million in 2020, at a CAGR of 52.7% between 2020 and 2025. The need to control drug development and discovery costs and time, increasing public and private investments in AI in genomics, and the adoption of AI solutions in precision medicine are driving the growth of this market. However, the lack of a skilled AI workforce and ambiguous regulatory guidelines for medical software are expected to restrain the market growth during the forecast period.

Browse in-depth TOC on "Artificial Intelligence (AI) in Genomics Market"141 Tables24 Figures154 Pages

Download PDF Brochure: https://www.marketsandmarkets.com/pdfdownloadNew.asp?id=36649899

List of Key Players in Artificial Intelligence in Genomics Industry:

Drivers, Restraints, Challenges and Opportunities in Artificial Intelligence in Genomics Industry:

Key Findings of Artificial Intelligence in Genomics Market Study:

Request Free Sample Pages: https://www.marketsandmarkets.com/requestsampleNew.asp?id=36649899

Based on offering, the AI in genomics market is segmented into software and services. The software and services segment accounted for largest share of the global artificial intelligence in genomics market in 2019. Software is needed to generate new insights from large-scale datasets and help understand genomic variations, thus enhancing the search for disease-causing variants and reducing clinical analysis times. The benefits offered by AI in software are driving its adoption among end users.

Based on functionality, the AI in genomics market is segmented into genome sequencing, gene editing, clinical workflows, and predictive genetic testing & preventive medicine. Genome sequencing was the largest functionality segment in this market in 2019 and is estimated to grow at highest CAGR in coming years. The large share of this segment can be attributed to the use of AI solutions to identify chromosomal disorders, dysmorphic syndromes, teratogenic disorders, and single-gene disorders.

Geographical Growth Scenario:

The global AI in Genomics market is segmented into North America, Asia Pacific, Europe, Rest of the World. North America (comprising the US, and Canada) is expected to account for the largest share of the global AI in Genomics market in 2020, followed by Europe. The large share of North America can be attributed to the increasing research funding and government initiatives for promoting precision medicine in the US.

Speak to Analyst: https://www.marketsandmarkets.com/speaktoanalystNew.asp?id=36649899

Browse Adjacent Markets:Healthcare IT Market Research Reports & Consulting

Browse Related Reports:

Read this article:
Artificial Intelligence in Genomics Market worth $1,671 million by 2025 says MarketsandMarkets - Benzinga

Reconciling the AI Value Chain with the EU’s Artificial Intelligence Act – CEPS

The EU Artificial Intelligence Act (AI Act), proposed by the European Commission in April 2021, is an ambitious and welcome attempt to develop rules for artificial intelligence, and to mitigate its risks. The current text, however, is based on a linear view of the AI value chain, in which one entity places a given AI system on the market and is made accountable for complying with the regulation whenever the system is considered high risk. In reality, the AI value chain can present itself in a wide variety of configurations. In this paper, in view of the many limitations of the Act, we propose a typology of the AI value chain featuring seven distinct scenarios, and discuss the possible treatment of each one under the AI Act. Moreover, we consider the specific case of general-purpose AI (GPAI) models and their possible inclusion in the scope of the AI Act, and offer six policy recommendations.

See more here:
Reconciling the AI Value Chain with the EU's Artificial Intelligence Act - CEPS

Artificial intelligence discovers the ‘toledano steel’ of the future – Morning Express

For millennia, humans have imposed themselves on nature or other humans by mastering the art of melting and mixing metals: the Copper Age was followed by the Bronze or Iron Age. Modern steel is at the base of the Industrial Revolution of the late 18th and 19th centuries. In the 20th century, aluminum alloys, titanium or superalloys allowed enormous technological leaps in cars, planes, missiles, prostheses In the second decade of this millennium, a machine has discovered several alloys that equal and even surpass those created by humans in some of its properties.

A group of researchers from prestigious European technical research centers, from the Max Planck Institute for Metallurgical Research to the Delft University of Technology, passing through the Royal Institute of Technology in Stockholm, have now created a machine learning system (machine learning , in English) capable of diving among millions of combinations between the different elements of the periodic table, finding 1,000 candidates with the properties that interested them and analyzing them looking for those that theoretically would have a low coefficient of thermal expansion (the expansion or contraction of the material with cold or heat). As published in the magazine Sciencefound four new alloys with a coefficient equal to or lower than the most temperature-immune combinations used so far.

Until a few years ago, an alloy was essentially a mix between a parent metal and small concentrations of other elements from the periodic table. The rules of metallurgy almost forbade going further. The director of IMDEA Materials, Jos Manuel Honrubia, exemplifies this by comparing a coffee with an alloy based on iron. By dissolving the sugar, you get a single liquid with properties different from those of coffee and sugar separately. In alloys it is similar, but there are limits to the proportion of other elements that you can add to iron before there are precipitates that are no longer part of the main alloy and generally worsen its properties. All of this was blown up in 2004: Then two independent groups combined five elements in similar proportions, seeing that they formed a single unique solution, he says. This opened a new era in materials science, that of high-entropy alloys. But there was a new challenge: finding new combinations between a main element and smaller quantities of two or three others (steel is iron with three or four additions) was a difficult task, but feasible. Before this time, the addition of many alloying elements in large proportions was a problem. In those of high entropy, the possible new compositions of dozens of elements and their different concentrations are estimated to exceed 10. An amount impossible for humans to handle, but less so for machines.

Compared to traditional methods, machine learning is much more efficient, saving time and effort

Ziyuan Rao, scientist at the Max Planck Institute for Metallurgical Research

The researcher at the Max Planck Institute and first author of the research, Ziyuan Rao comments on the main advantage of his artificial intelligence (AI) system: Compared to traditional methods, machine learning is much more efficient, saving time and effort , He says. For most of history, the discovery of new alloys with better properties has been based on trial and error, the knowledge accumulated by craftsmen or directly serendipity. This is the case of Toledo steel, whose swords were feared for centuries. As the director of the National Center for Metallurgical Research (CENIM-CSIC) Carlos Capdevila recalls, they forged them with charcoal from nearby mountains, which contained more carbon than other swords in Europe, giving them more hardness. Materials science currently relies on computer programs and models that save calculations and anticipate results, but the decisive work remains human.

Rao and his colleagues artificial intelligence system consists of three basic steps. They first use a model that generates new mixtures from a database that the researchers had previously assembled. This is because high-entropy alloys have a huge compositional spectrum and it is almost impossible to cover all possible compositions, he details. In a second step, they use another model to predict the properties of the compositions they obtained in the first. In a final step, the system scores the candidates (in this case 1,000) by combining the expected coefficient of each with their degree of novelty.

They thus arrived at four new alloys that they compared with invar. It is an alloy that, in its original mixture, had 64% iron, another 36% nickel and small amounts of manganese, carbon and chromium. Discovered at the end of the 19th century, whose discovery earned the Nobel Prize for its creator, the Swiss Charles douard Guillaume, it had a very low coefficient of thermal expansion. Not being affected by thermal changes, it was and still is essential in the design of precision instruments, clocks, pendulums, motor valves, mechanics of telescope optics Rao assures that two of the alloys created by their intelligence system equals invar alloys and two others have the lowest coefficient of thermal expansion of high or medium entropy alloys.

Stefan Bauer, a researcher at the Royal Institute of Technology in Stockholm and one of the senior authors of this research, recalls in a note: Machine learning models have been incredibly successful when unlimited amounts of data are available, for example in video games. . However, in the real world, it is much more difficult to find use cases where artificial intelligence makes a difference. It is very exciting to see that the predictions were not only tested in simulations, but that new alloys were created and physically demonstrated. Having proven its worth with thermal expansion, the scientists intend to use their machine learning system to investigate other properties, such as magnetism, in other materials.

Jon Mikel Snchez is a researcher in advanced materials at Tecnalia. A few years ago he did his doctoral thesis on high-entropy alloys. When he is asked about the possible properties beyond thermal expansion of these alloys and their possible applications, he almost runs out of paper. There are so many alloys that have improved the traditional ones in many aspects. Some scientists compare the discovery of it with that of steels. Some have better magneto-thermal properties. Others have better cryogenic performance, key for fuel storage. He also recalls a high-entropy titanium alloy that outperforms the best titanium alloy used in prosthetics today. Lastly, one of the most important and the one that mortals understand best, better structural properties (vehicle parts, for example) especially at high temperatures. Hence, Snchez believes, the relevance of these works. Applying AI to discover new alloys is quite new. Discovering new materials by these methods is a significant advance, he says.

Capdevila, the director of CENIM, comments that discovering a new alloy or improving the properties of existing ones by slightly modifying their composition has its advantages. He gives the example of the cover that they are going to put on the Santiago Bernabu soccer field. Stainless steels have a high reflectance and without modifying them, the temperature on the surrounding terraces would be very high. However, the alloy they will put in neutralizes most of the heat. Discovering a new alloy would be for a doctoral thesis of four or five years, now the machine does it in a few days. But Capdevila emphasizes that the human part is still there. Its computing power, but I, human, tell it what parameters interest me.

Torralba, the director of IMDEA Materials, is convinced that high entropy alloys are beginning a new era. They promise improvements in highly demanded properties, such as certain magnetic properties, high resistance to corrosion, greater tolerance to extreme temperatures or thermal changes and remember that one of the obstacles to the development of fusion energy is the lack of a material that can withstand the high temperatures generated in a fusion reactor. In all technologies, progress depends on the necessary materials being available, he recalls.

You can follow THE COUNTRY TECHNOLOGY in Facebook Y Twitter or sign up here to receive our weekly newsletter.

See original here:
Artificial intelligence discovers the 'toledano steel' of the future - Morning Express

Letters: Official information, Ukraine, Julian Assange, taxation, and trees – New Zealand Herald

Chief Ombudsman Peter Boshier. Photo / Supplied, File

It's our informationFollowing repeated complaints from respected journalists, the Chief Ombudsman has expressed his disquiet, and put public sector chief executives on notice concerning the dismissal of requests for material available in compliance with the Official Information Act. He said that such government officials should cease their obstruction, be held to account as a key performance indicator, and "must give acceptance morally of the law".These unusually reproachful words from such a highly respected and careful figure as the Ombudsman are very telling.This intervention is yet another manifestation of how out of control our once democratic society has become. It is perhaps also a warning to Jacinda Ardern that her Government has improper influence over senior public servants who we trust to be fully independent. Hylton Le Grice, Remuera.

Unwinnable warUS Secretary of State Antony Blinken's grand statement: "One man chose this war. One man can end it," rejects responsibility for the significant escalation of the war by the UK, US and Volodymyr Zelenskyy at the end of April, when they announced they would now fight to win, taking separatist territory and the Crimea. Now Russia is losing, Vladimir Putin has reacted with annexations and a nuclear threat. We need to get off this tit-for-tat train to nuclear war. Our leaders are acting as though there's no choice. There's always a choice to de-escalate from nuclear war, which is the insane option. How fast we got to this point is horrifying. If we stop fighting to win, serious negotiations can proceed. It doesn't matter who's done what. Our own Government should condemn all escalation and urge a war freeze.Real peace demonstrations by citizens would help: a demonstration urging one side to win is not a peace demonstration.S R Jacobs, Glenfield.

Brutal repetitionThe situation in Ukraine needs to heed the lessons of history, particularly the 20th century.First, when dealing with autocratic dictators, appeasement never works.Second, no matter how big your army or how hard you puff and blow, if the invaded country doesn't want you there, you can't win. Think Vietnam or Afghanistan.Tragically the Ukrainian situation will continue for some time. The lessons will ultimately be the same as we have already learned. The cost of the lessons? More blood and suffering for ordinary people.Richard Alspach, Dargaville.

Assange extraditionWe are seeing accelerated protests around the world against the UK Government's agreement to extradite Australian journalist Julian Assange to the US for publishing information on US war crimes and misdeeds in Iraq. Aotearoa-New Zealand groups have formed in support of Assange, who was reporting facts in the public interest. Aucklander Matt O Branain has inspired the idea of a human chain around the Houses of Parliament in Britain on October 8, an idea welcomed by Assange's wife. Political leaders and thousands of citizens are already joining this protest action.Assange is meanwhile detained in a high-security London prison, his death said to be imminent if he is not released from his solitary confinement there. Human rights groups, Amnesty International, Reporters Without Borders and the International Bar Association among others are demanding an end to the prosecution of this case. His extradition will, according to the UK's National Union of Journalists, seriously undermine press freedom and "chill the media worldwide".Jane Holst-Larkin, Grey Lynn.

Take and takeSo much for our "empathetic" Prime Minister and her Government who not only tax every dollar a person earns, regardless of how little they earn, but then turn around and tax them 15 per cent again for the food they buy to feed their family and any medications they need that Pharmac will not fund. It is beyond disgusting.Ericson List, Ppmoa Beach.

Left standingSo, I've watched the debates, read the articles, and now it comes to this.Who is going to speak out about the reckless cutting down of beautiful trees?Climate change is everything, getting out of cars is massively important but, as Aucklanders, aren't we just sick to death of our beautiful trees being annihilated? I include the threat on Mt Albert. At the end, it probably comes to this for me.Samantha Cunningham, Henderson.

Indecent proposalAuckland mayoral candidate Wayne Brown's vulgar outburst about respected Herald journalist Simon Wilson saying that, if elected, he would stick pictures of Wilson on urinals so people could pee on him does not represent many decent-minded people in Auckland and shows that Brown does not have the mettle required to be in politics.Wilson's critical analysis of Brown, whether one agrees with it or not, serves an important function of journalism, which is to shine a light on those who seek power.As the saying goes - if you can't take the heat, get out of the kitchen.Raewyn Maybury, Westmere.

Safe environmentYou have to admire Karen Chhour (NZ Herald, September 30). What a caring, sensible great Kiwi.Karen is wanting to make sure children are in a loving, safe environment. Unfortunately, this section 7AA has been around for some time and, as she said, children have been put back into an unsafe environment with abuse and violence.

I thought that whnau were meant to be caring, loving people. Yes, I am a whitey but have principles and respect others, whether they are white, black yellow or green.Good on you Karen, I don't think I could accept the apology that you have.Hopefully, you have better luck getting action on Section 7AAVickie Corbin, Mahinepua.

Bullying formsI wish to share some of my observations as a primary, intermediate and high school teacher of 54 years. As intermediate teachers, we were required to teach "kia kaha", anti-bullying. There are some underhand forms of bullying, I observed, that I believe can lead to violence and which, I've observed, can be learned from approximately 2 years old.Examples are, nag/whine/attention-seek; poor me/victim; dump guilt. To the latter, "I don't accept that guilt," spoken softly, is a very effective reply.Other examples are repetitively using "I'm so cute!" e.g. often daddy's darling, including using "big eyes"; to manipulate a situation; go vague; bulldoze/hammer e.g. talk nonstop; produce crocodile tears; tease; ignore problem e.g. walk away/run away/flounce off; throw a tantrum/storm off; use blackmail; repetitively use sarcasm/ mockery/ derision; inappropriately laughing; not talking, for a long time, known as the silent treatment.If I mentioned a tactic - for example, dumping guilt - to the children in my lovely class, they would turn and stare at a child who often dumped guilt.Kiri O'Neill, Cambridge.

Well managedCongratulations to the Government for creating the conditions that mean that every successful business is now doing better than it has ever done. Even Bloomberg congratulates the Reserve Bank for moving early in the current tightening cycle. We can easily see the catastrophic consequences of National Party proposals for the economy by looking at the UK. Labour has always been the better economic manager and today's economy was both predictable and predicted.Mark Nixon, Remuera.

Over the humpIn addition to fixing the potholes, now that speed limits on Auckland roads have been reduced to little more than walking pace, how about Auckland Transport removing all the speed humps and raised pedestrian crossings throughout the city and perhaps bring back the men with red flags preceding motor vehicles in lieu? Possibly the new mayor could facilitate this.J G Olesen, St Heliers Bay.

On UkraineWho has killed the greater portion of people and done the greater portion of the physical damage? Has Russia actually been killing those that it claims are Russian citizens? Mike Wells, Kawerau.

On squareThe newly-renovated CPO building looks very fine. However, I note with great sadness that the open space in front is no longer called Queen Elizabeth Square. John Hampson, Meadowbank.

On ChhourKaren Chhour (NZH, Sept. 30) writes with love and respect, something lacking in any of Willie Jackson's comments and attitude. S. Hansen, Hastings.

On votingMaybe we don't engage in local politics because, simply, we struggle to see value in what they do. John Ford, Taradale.

On violenceDeliberate acts of violence whether in a rugby match or anywhere else should not be tolerated and the perpetrator should be facing criminal charges. Bob van Ruyssevelt, Glendene.

On roadsReducing the speed will not achieve "Road to Zero", without putting higher penalties on non-use of seatbelts, use of phones and devices while driving, and driving whilst under the influence of alcohol and drugs, for a start. Marie Kaire, Whangrei.

On Vodafone"O" is for "Awful". Martin Adlington, Browns Bay.

Liam Dann: Brace for an action-packed inflation reveal

One thing is for sure; if inflation drops below the present figure (unlikely given my supermarket and fuel bills) it will all be a result of "prudent fiscal management by this Labour Government"; if inflation remains at the present figure or exceeds it (likely given my supermarket and fuel bills) it will all be a result of "international pressures and a Covid overhang, but nothing to do with the internal management of our booming economy by this Labour Government". Beware, statistics can be made to tell whatever story is desirable; reality of statistics is the increased prices you are paying. Andrew R.

Do you realise you are accusing the Government of spin and bias, then doing exactly the same? Your accusation may be accurate, but at least acknowledge you are equally one-eyed. Susan H.

We just want the truth, not spin, not vanilla. Meanwhile, the odd tongue-in-cheek comment does no harm. Jim S.

On the flip side though, if inflation continues, National will be trumpeting that it's fully down to Labour's mismanagement and money printing. If inflation reduces, they'll be saying it's in line with global inflation easing and has nothing to do with the Government's management. They're all playing the same game. Harry W.

Yes, politics is politics but that doesn't excuse the current Government's wasteful spending, nothing ever does. And in 2022 we have the war but, if you look at 2021, our inflation was among the worst. Sudhir M.

Whatever figure drops out it will not reflect the reality of the everyday costs to most New Zealanders struggling to keep an even keel, let alone save, except perhaps those at the top of the tree in pay terms, e.g. politicians and bureaucrats. Garry P.

See original here:

Letters: Official information, Ukraine, Julian Assange, taxation, and trees - New Zealand Herald

Discussing the death and life of Seth Rich with Andy Kroll: podcast and transcript – MSNBC

Seth Rich was a young DNC staffer in Washington who was tragically murdered early one morning in 2016. Our WITHpod guest this week described him as smart, ambitious, telegenic and someone who might run a presidential campaign someday. In the absence of an arrest, questions remain about who killed Rich. Unfounded theories about the motives for his murder continue to circulate on social media, including ones that enmeshed the Clintons and other high-profile figures. The search for answers, and this age of widespread disinformation, is the subject of A Death on W Street: The Murder of Seth Rich and the Age of Conspiracy, written by ProPublica reporter Andy Kroll. The true-crime story unravels this saga of murder, deceptions about what happened and the role of conspiracy mongers in disparaging Richs memory. Kroll, who actually knew Rich, joins WITHpod to discuss Richs life, death and what happened to his story once it got into the hands of numerous bad actors.

Note: This is a rough transcript please excuse any typos.

Andy Kroll: Someone you met at a party one time and thought to yourself, man that guy is going somewhere. He's smart, he's telegenic, he's ambitious. He's going to be running a presidential campaign in the next 10 years, no question. And then you find out he's dead, murdered not that far from where I live here in D.C.

So, I felt it on that personal level, but then when this local news story, this private family matter gets transformed into a political story, a viral meme, a hashtag, a billion threads on Reddit and 4chan, that's when the switch happened for me. That is when I felt these two separate worlds of mine collide, the personal, the day job and I thought I just have to know what the heck is going on here.

Chris Hayes: Hello and welcome to the "Why Is This Happening?" with me, your host, Chris Hayes.

You know, I think one of the central experiences of our age is a sense of constant vertigo and dislocation as regards information about the world. First of all, there's just a lot of it, there's obviously too much of it to pay attention to. There's also a lot of things that are just wrong that are floating around. It's very hard to figure out what's wrong and what's right, sometimes a tweet will go viral, and it will turn out to be like satire or photoshopped or some random person like took something wildly out of context.

And then at a bigger level, you see entire media platforms devoted to untruths, whether that's about the election lie or about vaccine efficacy. And I mean, look, it's easy to get overly presentist about this. It's always hard to separate fact from fiction and things that are true from things that are not. The world is complicated and there's all kinds of stuff, like I love when there's like some big dispute in some country you don't follow, and you like you try to get into it. It's like, was the trial against Lula in Brazil like actually corrupt or like did he do the thing? And it's like, well good luck trying to figure that out, just come beaming in from 30,000 feet, particularly when you arrived in very contested debates.

That said, if you were to ask me, what's the moment where it felt like we veered off into a new level of surreality, disinformation, confusion, and vertigo, I think it's pretty clearly 2016. Like, the 2016 was really -- it really did feel like that year and that campaign and the rise of Donald Trump represented us moving off course that we were on. Not the course we were on was like amazing, like, well everyone agreed about the facts like, you know, like one of the major parties had been denying climate change for 20 years. But the level, the acuteness of almost deranged counterfactual narrative and disinformation was truly headspinning. It elevated to the highest levels. It moved from the margins to the center. It set the agenda often for mainstream discourse.

And there's one particular example of that, that is in some ways a microcosm, in some ways a kind of allegory, and in some ways just an actual example of this phenomenon, which is the death of a DNC staff named Seth Rich. Seth Rich was a young, DNC staffer who tragically was murdered in Washington D.C. late one night in 2016. A private tragedy, an awful, awful, awful thing to happen, brutal for his family and friends and the people who love him, that then got pulled into an updraft of conspiratorial insanity that basically pointed to him as a central figure in this grand conspiracy.

My guest today is someone who wrote a book about this chronicle, about Seth Rich, his life, his death, his legacy, and what happened to his story once it got into the hands of all kinds of bad actors. It's an incredibly well-told tale. It's very humane and empathetic and also really provocative and an incredible tale about what it means to live and die in the informational universe we live in now. His name is Andy Kroll. He's a reporter of ProPublica. I've known him for a long time. He's a great reporter, and the book is called, "A Death on W Street: The Murder of Seth Rich and the Age of Conspiracy." And Andy it's great to have you in the program.

Andy Kroll: It's great to be here. Thanks for having me, Chris.

Chris Hayes: Just talk to me first about the origins of this because you knew Seth Rich. This starts as a personal story, not a reporting story. So, just tell us a little bit about who he was, how you knew him and what happened.

Andy Kroll: It was a strange experience for me as a reporter. When you start on a big story, let alone a book, it comes from a source that you have or a particularly intriguing piece of information you read in the news, a tip that comes across the transom, and you take it in, absorb it and pursue it with your investigative reporter hat on.

But for this story, it was a text message from a friend of mine, someone who has nothing to do with journalism, nothing to do with politics, anything like that, just a buddy, with a link to the local news story that said Seth had been killed on July 10th, 2016 in this tragic matter of a wrong place, wrong time situation.

And for a couple of weeks there, I followed the news of what had happened to Seth as the details trickled out, as memorials happened and the funeral took place back in his hometown of Omaha, as not a journalist but a peer or someone who traveled in similar social circles. It's like someone you met at a party one time and thought to yourself, man, that guy is going somewhere.

Chris Hayes: Yes.

Andy Kroll: He's smart, he's telegenic, he's ambitious. He's going to be running a presidential campaign in the next 10 years, no question. And then you find out he's dead, murdered not that far from where I live here in D.C., so I felt it on that personal level.

But then, when this local news story, this private family matter gets transformed into a political story, a viral meme, a hashtag, a billion threads on Reddit and 4chan, that's when the switch happened for me. That is when I felt these two separate worlds of mine collide, the personal, the day job, and I thought, I just have to know what the heck is going on here. I have to know how this has happened.

There is this moment when the personal and the journalistic collided. I remember sitting at my desk in August, probably, of 2016 and I saw #sethrich trending on Twitter. I thought to myself, what the heck is going on? How is that possible? I had followed some of the small conspiratorial chatter that had popped up right after he'd been killed, but not at this level, nothing had blown up in the way that I was now seeing it blow right before my eyes. And really from that point onward, August of 2016, I've been chasing the story, reporting on it, trying to understand again what happened, how this could happen, and eventually got to a point where I thought, I can't fit this into a story or two or three. This is really a book, and that's what led me to write the book.

Chris Hayes: Tell me about who Seth Rich was. What was he doing for the DNC and what do we know about the circumstances of his death?

Andy Kroll: Seth was from Omaha, Nebraska, a Democrat in an overwhelmingly Republican state. He grew up weaned on The West Wing, on watching C-SPAN in his free time in his bedroom at home. He was a total political nerd, a junkie. He followed Congressional races and redistricting fights in his home state the way we follow sports scores and eventually how our college did on a football game over the weekend. He was obsessed with this stuff.

He moved to Washington the first chance he got after graduating from Creighton University again in Omaha, and he wanted to be in the middle of the action. He was like so many people who flocked to Washington after college. They want to make their mark. They want to make a difference in the world. They want to play some small part and maybe someday a bigger part in the story of the country and its government and that was Seth.

He described himself as a patriot. He wore crazy stars and stripes outfits on 4th of July, in part as a sort of winking gag with his friends, but in part also because he believed that stuff. He was earnest about how much he loved his country, how much he cared about American democracy. When he was killed, he was working for the DNC in the voter expansion department. He was the only non-lawyer on the team of lawyers, trying to figure out ways to expand the franchise basically, how do we find Democratic votes, wherever they are, and get them to vote? How do we find people who aren't registered, get them to register, so that they can vote?

He really believed in voting is the lifeblood of the country and that regardless of whether you're a Democrat, you're a Republican, the country was at its best when everyone was participating, everyone is voting, everyone's voice was heard. That was what he was doing on the day he was killed. He was about to accept the job on the Clinton campaign, doing similar work, and that would have fulfilled a dream of his. He always wanted to work on a presidential campaign and he was maybe a week or two away from that when he died.

Chris Hayes: How old was he?

Andy Kroll: Twenty-seven years old.

Chris Hayes: And he had been at the DNC for a few years at that point?

Andy Kroll: Yes, he had been there for two and a half, three years at that point. He had been through a bruising midterm election, which I think opened his eyes to the less savory parts, the less glamorous parts of working in politics. I think he had come to see that politics in real life is not The West Wing and not everyone is walk and talks and quippy one-liners and the idealism of President Bartlet.

He loved that show but was also coming to grips with the fact that, that's not exactly how politics works, certainly not how politics worked in 2016, as you described earlier/ But he still wanted to work in voting. He wanted to continue this passion of his, something again that he had been passionate about since he was in high school.

Chris Hayes: So, he's coming home from a bar one night, which is something that I've done in Washington D.C. I have to say that someone I knew, Brian Beutler who now is at Crooked was shot and was in critical condition under extremely similar circumstances back when I was living in D.C. You know, it was someone I knew. They were walking down a block I'd walked down. They had come back late at night. Brian survived, it was extremely traumatic. He's written about it.

But this sounds like a somewhat similar situation basically, like random street crime as far as we can figure.

Andy Kroll: I remember hearing when Brian was shot all those years ago, and I thought about it when I first heard about Seth, walking home 2:00, 3:00 in the morning anywhere in D.C. can be a problematic situation through no fault of the person's own. Seth lived in a neighborhood in D.C. called Bloomingdale, Northwest, but just barely so, kind of hugs North Capitol Street, the big North-South Corridor here.

Bloomingdale at the time had been plagued by armed robberies all summer long. Weirdly same MO as well, two guys, one with a gun, robbing people for their valuables, especially their iPhones, and you can see this in the police reports, which I pulled and compiled for the book. The two guys would stop people, usually people who were talking on the phone, which Seth was doing at the time he was killed. They would ask for the phone but they would say, disable the Find My Phone tracking app, and then gives us your phone with the gun pointed --

Chris Hayes: Right.

Andy Kroll: -- at their face. And the police have said all along that Seth's murder and the circumstances of it certainly sound a lot like all of these other armed robberies that were happening in Bloomingdale, in Seth's neighborhood. The neighborhood had also been ripped up because they were putting a tunnel underneath it. It floods all the time, so they're trying to fix this problem. And so it's kind of an open-air maze, there's fences everywhere, lights were knocked out. It was just not a good place to be walking around that late at night, talking on the phone.

And from the day that this happened, and from the day that the police announced that Seth had been killed, they said this was an attempted robbery, maybe Seth tried to fight back, maybe there was some kind of altercation. There were some markings on him that suggested that there were, and he was shot several times and did not survive that attack, which unfortunately happens all the time in major American cities. But the details of this crime would fuel so many of the conspiracies that would come afterward, in part because people were looking for a reason to doubt. They weren't looking to buy the official story from the police.

Chris Hayes: Well and it also was a homicide that was not solved. I mean it was not cleared at the time that when the conspiracy theories take off, but that too was extremely common in major American cities and D.C., sort of a shocking percentage of homicides go unsolved.

Andy Kroll: Right, right, that's right. And this is unfortunately one of those cases, the murder is unsolved to this day. The investigation is active but now we're going on six and a half years or so, six-plus years. The police haven't announced suspects. They haven't announced any leads yet, even though the investigation is active. And that alone fuels these theories that come --

Chris Hayes: Right.

Andy Kroll: -- afterward.

Chris Hayes: Right, and part of my point there is that I think part of that is people's unfamiliarity with how shockingly and awfully common it is for people in the West Side of Chicago, in Anacostia Washington D.C., in all kinds of neighborhoods throughout America, particularly neighborhoods that are poor and predominantly non-white, that murders happened without being solved. That happens quite a bit.

When I think I've seen, you know, when I've seen this sort of conspiracy theories like this extra air of mystery that is unsolved, it's like, yeah, a lot of murders in America are unsolved. So, this happens. It's a horrible tragedy obviously and profoundly upsetting for his family, for his friends from around him.

What are the first inklings that this is going to move from a private and terrible tragedy to something else?

Andy Kroll: It's remarkable how fast those initial inklings appear. I went back and almost like a social media archeologist of sorts, retraced as best I could, the origins of these theories about Seth. What I found really interesting was that they began on the far left end of the political spectrum before they eventually moved and would really take off on the right end of the political spectrum.

And you got to go back to, again, this chaotic, insane year of 2016, this presidential election like no other. In the summer of 2016, one of the biggest stories was the state of the Democratic Party and the near civil war between supporters of Senator Bernie Sanders, more on the progressive side, and supporters of Secretary Hillary Clinton, the more centrist establishment Democratic camp.

Clinton had just about sewn up the nomination. She would be named, crowned the nominee at the convention a couple of weeks after Seth was killed. But the animosity, the tension within the party at the time that Seth was killed was at a fever pitch. That is where the Seth Rich conspiracy theories first took hold. It took hold among Sanders supporters and supporters of then Green Party candidate, Jill Stein, to give you a real throwback, shout out here.

And there was speculation among these folks that Seth had been a whistleblower of some kind, that he had been trying to expose the DNC's wrongdoing as it related to Bernie Sanders, that he was somehow a Sanders supporter who is going to blow the lid on how the DNC had stolen the nomination or rigged the nomination process for Clinton and against Sanders, and that's the origin.

People forget that. It's easy to forget it because there's really only a window of a couple of weeks there before these theories would explode on the opposite end of the spectrum but that is where it started within the party. People who were thinking that Seth was this Bernie bro, for a lack of a better way to put it, who was angry about what he had seen in terms of the treatment by the DNC of Sanders.

Chris Hayes: And not just online, not just in threads and Twitter and those kinds of places the people are saying and the implications that it was a hit, right, that he's killed to keep him quiet, to stop him from spilling the beans about DNC wrongdoing.

Andy Kroll: That the Clintons or their emissary, someone working on behalf of the Clinton family had ordered a hit on this DNC staffer because he had tried to expose fraud, wrongdoing, nefarious backdoor dealings of the DNC, that's exactly right. You see this stuff popped up on Twitter. I mean I found tweets and some of them are still there to this day, within minutes of the official announcement, the news breaking on Monday, July 11th, that Seth had been killed. I mean there was no sort of period for percolating or --

Chris Hayes: Right.

Andy Kroll: -- moments where people are trying to figure out what do we say. I mean this was a reflex. This was almost immediate on Twitter, on Reddit, anywhere where there were sort of congregations of Sanders and Stein supporters.

Chris Hayes: More of our conversation after this quick break.

Chris Hayes: To the best of your knowledge, this is just random and essentially organic, right? I mean these are just people who are in an extremely paranoid mindset. There's of course like lineage of the conspiracy theory against the Clintons that goes back to Vince Foster's death and the Clinton body count. And the idea, you know, Vince Foster of course, a friend of Bill and Hillary who worked in the White House which occasion (ph) all kinds of crazy conspiracy theories that went very mainstream, particularly in Republican Party that he had been murdered. He had been whacked by the Clintons.

And there's this idea of like a Clinton body count where the Clintons just go around like having people killed. This was in the ether before Seth Rich and I think it's sort of a necessary context to understanding why anyone is making this particular leap.

Andy Kroll: The only part of the book where I stepped out of the main timeline, where I --

Chris Hayes: Yes, exactly.

Andy Kroll: -- jumped backward from the blow by blow on the book is this really sort of fascinating moment where it's a few days after Seth had been killed, the funeral just happened. Hillary and Bernie are about to appear on stage for their first sort of moment of reconciliation after Clinton had won the nomination effectively. Seth's old colleague, the DNC, thought to themselves, well hey, wouldn't it be great if Hillary said something at this event to remember Seth. Everyone is going to be watching. Media is descending from around the world on little ports of New Hampshire.

Can she like work Seth's name in the speech? She does, and I quote someone who had worked with Seth at the DNC saying, we watched Hillary Clinton say this and the gratitude that this person felt almost immediately melted away into a oh my God, what have we done, knowing that in the ether there's this long history, a lineage is a great word for it of Clinton conspiracy theories and that's the chapter in the book where I jumped back. I actually found this story that I haven't even known about going to the book, about a woman with intern in the Clinton White House.

Chris Hayes: Yup.

Andy Kroll: Very briefly, it was Caity Mahoney. She was an intern. She worked in like the tour guide office and then had left. She was working at Starbucks in Georgetown, a tiny neighborhood here in D.C. and was killed in an attempted armed robbery, noting was taken. She and two other colleagues were pretty brutally murdered by a robber, and then for several years the murder went unsolved, and she became the latest addition to this Clinton body count, that's next to Vince Foster, next to Ron Brown, the former Commerce Secretary, next to all of these people from Arkansas that no one had heard of but had somehow been attached to this Clinton body count list.

And I did that in the book because I feel like people needed to know. Why would folks on the internet immediately jump to the Clintons did this? Why would they immediately suspect a political hit job? I mean that is quite an accusation to make and yet people had been making it, the American politics --

Chris Hayes: Exactly.

Andy Kroll: -- 20 (ph) years at that point.

Chris Hayes: One of the things that you start to see there and you do this in that chapter is you start to see the sort of magnetic logic of conspiracy which is once you start looking for it, once the way you reasoning is, let's find anyone who had any brush of the Clintons who died an untimely death. You start to see this pattern emerged. Now the pattern is nonsense because the pattern could emerge for anyone.

That was how you started to look into someone, right? You're reasoning backwards. It's a deficient means of understanding the world and yet it does have kind of magnetic draw. And so with this already established, there's this kind of vortex pull, right, when Seth Rich dies, that this is a ready-made story there that people picked up immediately.

Andy Kroll: And then it gets to the underlying appeal of conspiracy theories writ large. You're confronted with an event that you find confusing, you don't understand, you're skeptical of what the people in power are telling you happened in this particular instance. Obviously, with trust in any institution on a constant decline, it seems like people are prime to doubt what the official story is. And so, they go looking for alternative explanations. They "do their own research" as we like to hear from conspiracy theorist all the time.

And in this case, it's so easy to make the leap from he was shot and killed, it was an armed robbery gone wrong to no, no, no to the nation's capital, this guy worked for the crooked DNC as so many people like to say it at that point in time. The Clintons have a history of doing this. Of course, something more was going on here. And you know you said a second ago, Chris, presumably this was organic. This was real people doing this.

You know, I found some data in the course of reporting a book showing that, yes, there was some back (ph) activity or the Russian Internet Research Agency came along a month or two after Seth was killed and they amplified some stuff. But this was not a creation of the Russians or the Chinese or some domestic troll farm. I've interviewed people. I interview them for the book who were among the earliest to say, this was a hit. This was politically motivated.

Some of them interestingly have backed away from it in the time since, like I quote a long Reddit thread that was posted again within hours of Seth's murder by a young man in Florida. And I tracked him down and talked to him, and in that case, he said I was just, again, trying to do my own research. I didn't necessarily trust mainstream media. He was a Sanders supporter. He thought that he could advance a story in a way that felt authentic to him.

And you know, five years later, four years later whenever I talked to him, he felt some regret about that, when I talked to him. But I have also talked to people who to this day say, nope, definitely it was a hit, way too many questions, doesn't add up, don't believe the police, you know, your book is nonsense.

Chris Hayes: Right. And I think this point those interviews are fascinating by the way and I was so glad you did them because I do think there's a little bit of a comforting fiction we tell ourselves about disinformation being some product of Russian interference. And clearly it is something that they have pushed and something they amplified, and it isn't a very effective means of like messing with the population but there's just a massive organic part of it too which is like people believe crazy stuff.

They get together in the internet and they goad themselves into believing crazier and crazier stuff and then in the case of where (ph) going to get to, it can get amplified, people with really big platforms. So, all of these is happening immediately after, it's working on the already well-established universe of Clinton conspiracy theories, about them as essentially serial murders who have like hit squads. This is July 10th, July 11th news. When is the first WikiLeaks?

Andy Kroll: August 10th, 2016, a month later.

Chris Hayes: So, it's a month later and that happens right before the DNC. It's timed very obviously to kind of like blow up the DNC, and I remember that because I was in Philadelphia, and it goes off. The first one happens right when, the first posting of WikiLeaks of the purloined e-mails of the campaign manager, John Podesta, right, are posted by WikiLeaks right when Trump is facing the worst crisis of his campaign which is the leaked Access Hollywood tape which he says you can just grab them by the P-word.

It's really disgusting, huge condemnation, people are fleeing him. This is obviously timed to counteract that and it's also right before the convention. What does the appearance of WikiLeaks and the WikiLeaks' e-mails do the Seth Rich conspiracy theorizing?

Andy Kroll: If there was a single moment that happens in the larger arc of the story that if it didn't happen would change the course of history, at least as it relates to Seth Rich and his family. WikiLeaks and Julian Assange's intervention would be that flashpoint. When you laid out this timeline, I'll add a couple of dots to it. You have WikiLeaks releases these stolen e-mails we now know taken from inside the DNC right before Philadelphia.

I was there too. I remember the look of sort of barely contained terror and fear in the eyes of most DNC employees I encountered in Philly. You have the release of Podesta e-mails right after Access Hollywood, very clearly intended to distract and deflect from that. And what you have in between those two things is this really critical moment in the Seth Rich story, critical moment in the book. Julian Assange is giving an interview to a Dutch TV station. To the interviewer's credit, he's pressing Assange, where did you get these stolen e-mails at the time, the DNC e-mail specifically.

You know, cyber security experts are saying this is most likely Russian-relate that some kind of hacking group affiliated with Russia took them and gave them to you. Assange is denying, he's deflecting, he's disingenuous and in this Dutch interview, he says, well, don't you know, he says this out of nowhere without prompting, without any suggestions. So, clearly, he had it in his mind. He says, well, there was this young DNC staffer who was murdered in Washington D.C. recently. Our source get concerned when they see things like that.

Chris Hayes: Oh so despicable.

Andy Kroll: Yes.

Chris Hayes: It's just so despicable. It is an unbelievably despicable thing to do.

Andy Kroll: Yes and I had never thought about Julian Assange and WikiLeaks the same since this moment. Obviously --

Chris Hayes: Yes, same here.

Andy Kroll: -- Unique have written a whole book about this, but it's clear what Assange is doing there but the effect of that online is an explosion of tweets, Reddit posts, memes, Instagram post, everything under the sun saying, Julian Assange just pointed a finger at this guy, Seth Rich. This Russia story is nonsense.

Chris Hayes: The Russian story is nonsense, right. So, there's a few things going on here, right? The DNC leak happens. When do they first start posting the DNC e-mails?

Andy Kroll: Right before the DNC convention, so that's like late July 2016.

Chris Hayes: Right. And it's after those e-mails start to come out that he gives his interview and --

Andy Kroll: Correct.

Chris Hayes: -- at that point there's already people inside the DNC who are like we know, you know, they had hired CrowdStrike, they had hired other outside firms to look at there, like we know this was foreign penetration, like we have a pretty good sense it's the Russians. These are not and those people doing that are not like FBI or government officials. These are just people who don't have a dog in the fight, right? They're brought into like basically run an audit and they're like, "Yeah, we kind of had figured it out where this came from."

And so, it becomes pretty clear that like the Russians have had to be in CE as a form of political interference. They've somehow gotten into Assange, and Assange is now publishing this which is completing the operation on behalf of the Russians whether he knows that or not. We don't know, right, if he knows where it comes from so I'm just going to give it to him.

And he goes on, this Dutch TV, to deflect attention away from the Russians and put it on Seth Rich and in doing so, A. Implying that he was murdered; B. Also implying that Seth Rich was this kind of whistleblower who had secretly been a source and then finally doing the thing that, if they were true, is the most irresponsible thing that you could possibly do for any journalist or anyone which is protecting sources. Like, you would never in a billion years say anything about who your source is if they give you something important.

So, all of that together, it just like the peak of disingenuousness and it's disgusting and that you say in the book, that's the moment where it blows up. It goes from like people saying on the internet to like a thing.

Andy Kroll: And I retraced the hours and days after the Assange interview because it was a little case study of how this tantalizing piece of disingenuous spin could go from the mouth of Julian Assange to the front of the Charge Report which obviously needs its firewall of the internet at that point, that becomes the subject of hundreds of thousands of tweets that ends up on Fox News the following night. You have Eric Bolling who was sitting in for Bill O'Reilly at the time just come out and say, plain as day, that Seth Rich had been killed in a hit. This was not an attempted robbery. This was a hit. That was Eric Bolling said on primetime Fox News after the Assange.

This is the first blow on the first Superspreader Event if you will of the Seth Rich conspiracy theory.

Chris Hayes: And Bolling says it's on Fox News. This is the first time that it gets introduced into the like the cable news Fox-viewing audience, right?

Andy Kroll: That's right, yup.

Chris Hayes: And he says this, I mean, it's just a crazy thing to say. I mean obviously, it's Fox News, it's Eric Bolling so, you know, not surprising but just a, like to say that there was a hit against someone is like a really freaking serious accusation to make.

Andy Kroll: I mean, one thing I started to struggle with in the book, and I feel like I got lost in the contemporaneous coverage of the story, people are making accusations all the time related to Seth Rich and I was always having the sense of, do you understand how serious the thing you're saying actually is? Do you realize the gravity of what you are accusing someone of doing or claiming had happened?

Yes, Eric Bolling, people forget the Bolling comment because it happened so early and because the later Fox News involvement in the Rich story which is so much more of a scandal, so much more of a bluff but yes, during the 2016 campaign a Fox News host was saying this was a hit, that there was a political crime, the murder of a democratic staffer in the middle of campaign for some related crime or some attempt on the part of this guy, Seth Rich. You know, I think people forget about it but Fox had been all over this almost from the beginning.

Read the original:

Discussing the death and life of Seth Rich with Andy Kroll: podcast and transcript - MSNBC

Nurses vote on strike action in first-ever RCN ballot – Yahoo News UK

SWNS

A lost dog dubbed "wonder pup" walked himself home FIVE MILES across a busy city while his owner searched for him.

Pip wandered off while chasing squirrels on a walk with owner, Libby Bowles, 47.

And while she spent 90 minutes searching for him in Leigh Woods, in Bristol, the pup took to the streets, and strolled home.

He was caught on CCTV during his 4.6-mile walk home, which included a stroll across Clifton Suspension Bridge.

The pooch arrived home 20 minutes before Libby did - after she'd taken to local lost and found groups to track him down, on September 18.

Locals posted updates and CCTV grabs as he was spotted travelling across the city - outside his old home, in the park, and outside a local museum.

Pip, a pedenco, is a rabbit hunting hound rescued from Spain.

Now a therapy dog, he is well-known around Bristol because he sits in Libby's backpack as she cycles around the city.

Libby said: "The thing is, he's very calm and placid unless there's something furry to chase.

"He's run off before but he's always come back, so when he didn't I was quite worried.

"I spent an hour going up and down our walking route looking for him, and luckily ran into some friends who went round to the other side of the woods to see if they could find him.

"They actually did see him, but then at the last minute he zipped away from them under a fence."

His escape sparked a city-wide chase, and he was captured on CCTV in several places across the city, trotting along the pavement unaware of the search party.

Ms Bowles said: "At first I thought 'how on earth is he going to cross Bristol by himself?'

"But thankfully Pip has a good nose - he often takes me to his dog friends' houses on our walks.

"The dog community in Bristol is amazing, so I put him in one of the groups and I got constant updates of where he was seen.

"He went back to our old flat, past Bristol museum, literally all over Bristol.

"Eventually he was seen in the park near our house, so I breathed a sigh of relief because I knew he should be able to get home from there.

"I called our neighbours and they were all waiting for him when he got back.

"He apparently trotted round the corner fairly nonchalant. He had all his dog friends and lots of treats waiting for him."

Pip once belonged to a hunter in Spain.

After being found on the streets, he was rescued and adopted by Libby, who works in sustainability education.

Pip is now a therapy dog, and is part of a programme called Read2Dogs - where children can read to him rather than adults to boost their literacy skills.

Libby said: "I used to be a primary school teacher and I think it's such a valuable exercise, it has a profound effect on confidence in the classroom.

"I'm writing some books about Pip and his adventures, so kids can read to Pip about all his exciting stories."

Read more from the original source:

Nurses vote on strike action in first-ever RCN ballot - Yahoo News UK

Machine Learning | Google Developers

Stay organized with collections Save and categorize content based on your preferences. Foundational courses

The foundational courses cover machine learning fundamentals and core concepts.

We recommend taking them in the order below.

New

A course to help you map real-world problems to machine learning solutions.

The advanced courses teach tools and techniques for solving a variety of machine learning problems.

The courses are structured independently. Take them based on interest or problem domain.

Clustering is a key unsupervised machine learning strategy to associate related items.

Our guides offer simple step-by-step walkthroughs for solving common machine learning problems using best practices.

Become a better machine learning engineer by following these machine learning best practices used at Google.

This guide assists UXers, PMs, and developers in collaboratively working through AI design topics and questions.

This comprehensive guide provides a walkthrough to solving text classification problems using machine learning.

This guide describes the tricks that an expert data analyst uses to evaluate huge data sets in machine learning problems.

The glossary defines general machine learning terms.

[{ "type": "thumb-down", "id": "missingTheInformationINeed", "label":"Missing the information I need" },{ "type": "thumb-down", "id": "tooComplicatedTooManySteps", "label":"Too complicated / too many steps" },{ "type": "thumb-down", "id": "outOfDate", "label":"Out of date" },{ "type": "thumb-down", "id": "samplesCodeIssue", "label":"Samples / code issue" },{ "type": "thumb-down", "id": "otherDown", "label":"Other" }] [{ "type": "thumb-up", "id": "easyToUnderstand", "label":"Easy to understand" },{ "type": "thumb-up", "id": "solvedMyProblem", "label":"Solved my problem" },{ "type": "thumb-up", "id": "otherUp", "label":"Other" }]

Read more:
Machine Learning | Google Developers

Machine Learning in Oracle Database | Oracle

Oracle Machine Learning AutoML User Interface

A no-code user interface supporting AutoML on Oracle Autonomous Database to improve both data scientist productivity and non-expert user access to powerful in-database algorithms for classification and regression.

Accelerate machine learning modeling using Oracle Autonomous Database as a high performance computing platform with an R interface. Use Oracle Machine Learning Notebooks with R, Python, and SQL interpreters to develop machine learningbased solutions. Easily deploy user-defined R functions from SQL and REST APIs with data-parallel and task-parallel options.

Data scientists and other Python users accelerate machine learning modeling and solution deployment by using Oracle Autonomous Database as a high-performance computing platform with a Python interface. Built-in automated machine learning (AutoML) recommends relevant algorithms and features for each model, and performs automated model tuning. Together, these capabilities enhance user productivity, model accuracy, and scalability.

Data scientists and data analysts can use this drag-and-drop user interface to quickly build analytical workflows. Rapid model development and refinement allows users to discover hidden patterns, relationships, and insights in their data.

Read the rest here:
Machine Learning in Oracle Database | Oracle

Learning on the edge | MIT News | Massachusetts Institute of Technology – MIT News

Microcontrollers, miniature computers that can run simple commands, are the basis for billions of connected devices, from internet-of-things (IoT) devices to sensors in automobiles. But cheap, low-power microcontrollers have extremely limited memory and no operating system, making it challenging to train artificial intelligence models on edge devices that work independently from central computing resources.

Training a machine-learning model on an intelligent edge device allows it to adapt to new data and make better predictions. For instance, training a model on a smart keyboard could enable the keyboard to continually learn from the users writing. However, the training process requires so much memory that it is typically done using powerful computers at a data center, before the model is deployed on a device. This is more costly and raises privacy issues since user data must be sent to a central server.

To address this problem, researchers at MIT and the MIT-IBM Watson AI Lab developed a new technique that enables on-device training using less than a quarter of a megabyte of memory. Other training solutions designed for connected devices can use more than 500 megabytes of memory, greatly exceeding the 256-kilobyte capacity of most microcontrollers (there are 1,024 kilobytes in one megabyte).

The intelligent algorithms and framework the researchers developed reduce the amount of computation required to train a model, which makes the process faster and more memory efficient. Their technique can be used to train a machine-learning model on a microcontroller in a matter of minutes.

This technique also preserves privacy by keeping data on the device, which could be especially beneficial when data are sensitive, such as in medical applications. It also could enable customization of a model based on the needs of users. Moreover, the framework preserves or improves the accuracy of the model when compared to other training approaches.

Our study enables IoT devices to not only perform inference but also continuously update the AI models to newly collected data, paving the way for lifelong on-device learning. The low resource utilization makes deep learning more accessible and can have a broader reach, especially for low-power edge devices, says Song Han, an associate professor in the Department of Electrical Engineering and Computer Science (EECS), a member of the MIT-IBM Watson AI Lab, and senior author of the paper describing this innovation.

Joining Han on the paper are co-lead authors and EECS PhD students Ji Lin and Ligeng Zhu, as well as MIT postdocs Wei-Ming Chen and Wei-Chen Wang, and Chuang Gan, a principal research staff member at the MIT-IBM Watson AI Lab. The research will be presented at the Conference on Neural Information Processing Systems.

Han and his team previously addressed the memory and computational bottlenecks that exist when trying to run machine-learning models on tiny edge devices, as part of their TinyML initiative.

Lightweight training

A common type of machine-learning model is known as a neural network. Loosely based on the human brain, these models contain layers of interconnected nodes, or neurons, that process data to complete a task, such as recognizing people in photos. The model must be trained first, which involves showing it millions of examples so it can learn the task. As it learns, the model increases or decreases the strength of the connections between neurons, which are known as weights.

The model may undergo hundreds of updates as it learns, and the intermediate activations must be stored during each round. In a neural network, activation is the middle layers intermediate results. Because there may be millions of weights and activations, training a model requires much more memory than running a pre-trained model, Han explains.

Han and his collaborators employed two algorithmic solutions to make the training process more efficient and less memory-intensive. The first, known as sparse update, uses an algorithm that identifies the most important weights to update at each round of training. The algorithm starts freezing the weights one at a time until it sees the accuracy dip to a set threshold, then it stops. The remaining weights are updated, while the activations corresponding to the frozen weights dont need to be stored in memory.

Updating the whole model is very expensive because there are a lot of activations, so people tend to update only the last layer, but as you can imagine, this hurts the accuracy. For our method, we selectively update those important weights and make sure the accuracy is fully preserved, Han says.

Their second solution involves quantized training and simplifying the weights, which are typically 32 bits. An algorithm rounds the weights so they are only eight bits, through a process known as quantization, which cuts the amount of memory for both training and inference. Inference is the process of applying a model to a dataset and generating a prediction. Then the algorithm applies a technique called quantization-aware scaling (QAS), which acts like a multiplier to adjust the ratio between weight and gradient, to avoid any drop in accuracy that may come from quantized training.

The researchers developed a system, called a tiny training engine, that can run these algorithmic innovations on a simple microcontroller that lacks an operating system. This system changes the order of steps in the training process so more work is completed in the compilation stage, before the model is deployed on the edge device.

We push a lot of the computation, such as auto-differentiation and graph optimization, to compile time. We also aggressively prune the redundant operators to support sparse updates. Once at runtime, we have much less workload to do on the device, Han explains.

A successful speedup

Their optimization only required 157 kilobytes of memory to train a machine-learning model on a microcontroller, whereas other techniques designed for lightweight training would still need between 300 and 600 megabytes.

They tested their framework by training a computer vision model to detect people in images. After only 10 minutes of training, it learned to complete the task successfully. Their method was able to train a model more than 20 times faster than other approaches.

Now that they have demonstrated the success of these techniques for computer vision models, the researchers want to apply them to language models and different types of data, such as time-series data. At the same time, they want to use what theyve learned to shrink the size of larger models without sacrificing accuracy, which could help reduce the carbon footprint of training large-scale machine-learning models.

AI model adaptation/training on a device, especially on embedded controllers, is an open challenge. This research from MIT has not only successfully demonstrated the capabilities, but also opened up new possibilities for privacy-preserving device personalization in real-time, says Nilesh Jain, a principal engineer at Intel who was not involved with this work. Innovations in the publication have broader applicability and will ignite new systems-algorithm co-design research.

On-device learning is the next major advance we are working toward for the connected intelligent edge. Professor Song Hans group has shown great progress in demonstrating the effectiveness of edge devices for training, adds Jilei Hou, vice president and head of AI research at Qualcomm. Qualcomm has awarded his team an Innovation Fellowship for further innovation and advancement in this area.

This work is funded by the National Science Foundation, the MIT-IBM Watson AI Lab, the MIT AI Hardware Program, Amazon, Intel, Qualcomm, Ford Motor Company, and Google.

More:
Learning on the edge | MIT News | Massachusetts Institute of Technology - MIT News

Study: Few randomized clinical trials have been conducted for healthcare machine learning tools – Mobihealth News

A review of studies published in JAMA Network Open found few randomized clinical trials for medical machine learning algorithms, and researchers noted quality issues in many published trials they analyzed.

The review included 41 RCTs of machine learning interventions. It found 39% were published just last year, and more than half were conducted at single sites. Fifteen trials took place in the U.S., while 13 were conducted in China. Six studies were conducted in multiple countries.

Only 11 trials collected race and ethnicity data. Of those, a median of 21% of participants belonged to underrepresented minority groups.

None of the trials fully adhered to the Consolidated Standards of Reporting Trials Artificial Intelligence (CONSORT-AI), a set of guidelines developed for clinical trials evaluating medical interventions that include AI. Thirteen trials met at least eight of the 11 CONSORT-AI criteria.

Researchers noted some common reasons trials didn't meet these standards, including not assessing poor quality or unavailable input data, not analyzing performance errors and not including information about code or algorithm availability.

Using the Cochrane Risk of Bias tool for assessing potential bias in RCTs, the study also found overall risk of bias was high in the seven of the clinical trials.

"This systematic review found that despite the large number of medical machine learning-based algorithms in development, few RCTs for these technologies have been conducted. Among published RCTs, there was high variability in adherence to reporting standards and risk of bias and a lack of participants from underrepresented minority groups. These findings merit attention and should be considered in future RCT design and reporting," the study's authors wrote.

WHY IT MATTERS

The researchers said there were some limitations to their review. They looked at studies evaluating a machine learning tool that directly impacted clinical decision-makingso future research could look at a broader range of interventions, like those for workflow efficiency or patient stratification. The review also only assessed studies through October 2021, and more reviews would be necessary as new machine learning interventions are developed and studied.

However, the study's authors said their review demonstrated more high-quality RCTs of healthcare machine learning algorithms need to be conducted. Whilehundreds of machine-learning enabled devices have been approved by the FDA, the review suggests the vast majority didn't include an RCT.

"It is not practical to formally assess every potential iteration of a new technology through an RCT (eg, a machine learning algorithm used in a hospital system and then used for the same clinical scenario in another geographic location)," the researchers wrote.

"A baseline RCT of an intervention's efficacy would help to establish whether a new tool provides clinical utility and value. This baseline assessment could be followed by retrospective or prospective external validation studies to demonstrate how an intervention'sefficacy generalizes over time and across clinical settings."

Read the original here:
Study: Few randomized clinical trials have been conducted for healthcare machine learning tools - Mobihealth News