Machine learning is perhaps old hat by now, but what's never going to be old hat is cats. People just can't seem to get enough of them. Learning Factory is an Early Access game that released last month about building an automated factory that produces the things cats want to buy, then sells them. Your job is to keep the shelves stocked and the cats happyand earn money by selling at optimal prices.

By making offers to cats your factory can train up machine learning models that will then automatically adjust market prices to account for trends and the wallets of the cats in question. Rich cats want fancy expensive cat towers and food, while normal cats just want a good deal on a ball of yarn, and construction worker cats want raw materials. It's a neat concept that bears out pretty well in action: Do you want to make a huge, all-inclusive single machine learning model or instead focus on specific models tailored to each customer type?

Learning Factory has just released on Steam Early Access. It's not that complicated yet, with about six hours of gameplay for me, but there's a lot on the developer's roadmap. Luden.io's previous game, While True: Learn(), also focused on Machine Learning and catsbut from the angle of language rather than commerce. You can learn more about Learning Factory on the official website.

If you're into factory games, have you checked out Dyson Sphere Program yet?

More:
Here's an adorable factory game about machine learning and cats - PC Gamer

Scientists have made a major advance in harnessing machine learning to accelerate the design for better batteries. Instead of using machine learning just to speed up scientific analysis by looking for patterns in dataas typically donethe researchers combined it with knowledge gained from experiments and equations guided by physics to discover and explain a process that shortens the lifetimes of fast-charging lithium-ion batteries.

It was the first time this approachknown as scientific machine learninghas been applied to battery cycling, said Will Chueh, an associate professor at Stanford University and investigator with the Department of Energys SLAC National Accelerator Laboratory who led the study. He said the results overturn long-held assumptions about how lithium-ion batteries charge and discharge and give researchers a new set of rules for engineering longer-lasting batteries.

The research, reported in Nature Materials, is the latest result from a collaboration between Stanford, SLAC, the Massachusetts Institute of Technology and Toyota Research Institute (TRI). The goal is to bring together foundational research and industry know-how to develop a long-lived electric vehicle battery that can be charged in 10 minutes.

Battery technology is important for any type of electric powertrain. By understanding the fundamental reactions that occur within the battery we can extend its life, enable faster charging and ultimately design better battery materials. We look forward to building on this work through future experiments to achieve lower-cost, better-performing batteries.

Patrick Herring, senior research scientist for Toyota Research Institute

The new study builds on two previous advances where the group used more conventional forms of machine learning to accelerate both battery testing and the process of winnowing down many possible charging methods to find the ones that work best.

While these studies allowed researchers to make much faster progressreducing the time needed to determine battery lifetimes by 98%, for examplethey didnt reveal the underlying physics or chemistry that made some batteries last longer than others, as the latest study did.

Combining all three approaches could potentially slash the time needed to bring a new battery technology from the lab bench to the consumer by as much as two-thirds, Chueh said.

In this case, we are teaching the machine how to learn the physics of a new type of failure mechanism that could help us design better and safer fast-charging batteries. Fast charging is incredibly stressful and damaging to batteries, and solving this problem is key to expanding the nations fleet of electric vehicles as part of the overall strategy for fighting climate change.

Will Chueh

The team observed the behavior of cathode particles made of nickel, manganese and cobalt (NMC). Stanford postdoctoral researchers Stephen Dongmin Kang and Jungjin Park used X-rays from SLACs Stanford Synchrotron Radiation Lightsource to get an overall look at particles that were undergoing fast charging. Then they took particles to Lawrence Berkeley National Laboratorys Advanced Light Source to be examined with scanning X-ray transmission microscopy, which homes in on individual particles.

An animation shows two contrasting views of how electrode particles release their stored lithium ions during battery charging. Red particles are full of lithium and green ones are empty. Scientists had thought ions flowed out of all the particles at once and at roughly the same speed (left). But a new study by SLAC and Stanford researchers paints a different picture (right): Some particles release a lot of ions immediately and a fast clip, while others release ions slowly or not at all. This uneven pattern stresses the battery and reduces its lifetime. (Hongbo Zhao/MIT)

The data from those experiments, along with information from mathematical models of fast charging and equations that describe the chemistry and physics of the process, were incorporated into scientific machine learning algorithms.

Until now, scientists had assumed that the differences between particles were insignificant and that their ability to store and release ions was limited by how fast lithium could move inside the particles, Kang said. In this way of seeing things, lithium ions flow in and out of all the particles at the same time and at roughly the same speed.

But the new approach revealed that the particles themselves control how fast lithium ions move out of cathode particles when a battery charges, he said. Some particles immediately release a lot of their ions while others release very few or none at all. And the quick-to-release particles go on releasing ions at a faster rate than their neighborsa positive feedback, or rich get richer, effect that had not been identified before.

We now have a pictureliterally a movieof how lithium moves around inside the battery, and its very different than scientists and engineers thought it was. This uneven charging and discharging puts more stress on the electrodes and decreases their working lifetimes. Understanding this process on a fundamental level is an important step toward solving the fast charging problem.

Stephen Kang

The scientists say their new method has potential for improving the cost, storage capacity, durability and other important properties of batteries for a wide range of applications, from electric vehicles to laptops to large-scale storage of renewable energy on the grid.

This research was funded by Toyota Research Institute. The Stanford Synchrotron Radiation Lightsource and Advanced Light Source are DOE Office of Science user facilities, and work there was supported by the DOE Office of Science and the DOE Advanced Battery Materials Research Program.

Resources

Park, J., Zhao, H., Kang, S.D. et al. (2021) Fictitious phase separation in Li layered oxides driven by electro-autocatalysis. Nat. Mater. doi: 10.1038/s41563-021-00936-1

Read more from the original source:
SLAC, MIT, TRI researchers advance machine learning to accelerate battery development; insights on fast-charging - Green Car Congress

The AI Standards Working Group, co-chairedby IBM Watson Advertising and Nielsen, today released theguideArtificial Intelligence Use Cases and BestPractices for Marketing, which is intended to help executives, marketers, and technologists get the most from artificial intelligence (AI) and machine learning (ML). The announcement was made duringthe IAB ILM annual conference.

The guide -- which provides dozens of working examples -- draws directly from the real-world experience of the co-chairs as well aspublishers, agencies, and ad-tech companies. Itincludes nine use cases that span internal robotic process automation and data migration for agencies as well as AI use cases for creative,contextual, video, and more.

advertisement

advertisement

David Olesnevich, head of product at IBM Watson Advertising, and IAB AI Standards Working Group co-chair, believes the work and the guideprovide a road map through insights driven by conversations about real-world challenges faced daily by some of the industry's largest brands.

It felt natural toform this group, so we reached out to IAB to generate momentum, because with privacy forwardness, changes in identifiers, and government regulators we will need technology that canscale, he said. We believe thats AI.

AI was once deployed in advertising at the DSP level, which is very transactional, but now is beingused across the entire stack, including planning and the development of creatives.

Bias is one of the key topics the group will address, based on feedback fromindustry participants following todays release.

The reason why bias is scheduled to be the next topic of discussion is because it was on everyonesradar, he said, adding that it is an issue the group will address head on with other partners like Dentsu, GumGum, and Dun & Bradstreet.

EvenGoogle is using AI to create Cohorts.

Onegoal of the guide is to share different perspectives on how AI and ML could work, along with adoption, as well as to understand different views based on groups within the industry.

Feedback from C-suite agency executives suggests they are apprehensive about aligning AI and ML to their business priorities and objectives, how to ensure teams havethe skills they need to use the technology, and finding ways to implement solutions with a third-party partner or leveraging existing internal resources.

Despite theconcerns, C-suite agency executives do expect the technologies to create a competitive advantage, and to increase profit margins by reallocating manual effort against work by offloading criticalbusiness function, leaving the tedious tasks to machines.

Marketers have a different take on what the technologies can do, and the apprehensions that might keep themfrom using them.

Marketers look toward AI and ML to automate ad creation for addressable and non-addressable programmatic media formats, automatic media mix modelingbased on brand and performance metrics, and optimize audiences based on addressable and non-addressable programmatic media formats.

They are apprehensive about the biastheir efforts might create. When it comes to bias, they wonder whether a media vendor is using AI or ML to benefit them or their clients, and are concerned about the lack of human touch that resultsin issues at the account level, and will adoption help or hurt their career development.

Technologists are also concerned that the algorithms running the AI and MLcould be biased and assume certain behavior based only on race or gender. They also have concerns that AI-based systems might not perform fast enough or scale to their intended level, and that thedata to build a performant algorithm might not be available or affordable.

Go here to read the rest:
IAB Artificial Intelligence Group To Build Standards, With Focus On AI, Machine Learning, Bias 03/10/2021 - MediaPost Communications

SSM Health, a nonprofit with $8 billion in revenue, provides its communities with high-quality care for vulnerable populations. One of the most vulnerable populations is made up of patients with kidney disease.

THE PROBLEM

Kidney disease is complex because 90% of people with the disease do not know they have it until they need dialysis or a transplant. There is little disease education or preventive efforts in the initial stages, making chronic kidney disease expensive to treat. Patients typically wind up receiving lower outcomes and lower quality of life than physicians would like to see.

CKD and end-stage renal disease patients manage 15-20 medications daily and have multiple comorbid conditions, complicating treatment.

"Patients with kidney disease make up under 5% of our patient population, but account for more than 20% of our total costs," said Carter Dredge, chief transformation officer at SSM Health. "We needed the focus and expertise that our partner Strive Health delivers through predictive analytics and the care team to better support our most at-risk population.

"Across the broad primary care base, providers are seeing patients with a range of health concerns, and CKD often involves just five to 10 patients in their panel," he continued. "During each visit, PCPs have limited time to meet these complex needs, and CKD symptoms are subtle. Often, patients were under-diagnosed for advanced CKD."

SSM Health needed a focused solution that helpedpredict the best time to engage patients to optimize the patient experience, improve outcomes and lower costs.

"At SSM Health, as our core clinical teams build the main programs that encompass all our patients and interventions across multiple populations, partnering with Strive Health has delivered focused care for a particularly complex condition that connects to the larger innovation pipeline, aiding the move to more risk-based contracts by helping build the required care coordination and analytics programs for more specific patient cohorts," Dredge said.

PROPOSAL

Analytics can offer diagnostic assistance and guide treatment decisions. Combining data from several sources, including claims, clinical data, live feeds from health exchanges, dialysis machines and demographic information for social determinants of health, algorithms can predict adverse events, including kidney failure during a given time frame or a cardiology event.

"The program we developed with Strive Health delivers comprehensive clinical services for CKD and ESRD patients that significantly improve quality of care and outcomes while lowering the total cost of care for patients," Dredge said. "Thirty-three algorithms assist with treating CKD, including one that can predict CKD progression to ESRD with 95% accuracy."

Carter Dredge, SSM Health

Strive Health's technology and full clinical model bring a focused approach to care, he added.

"We are intervening with the right patients at the right time," he explained. "Our care team can see when a patient is progressing more rapidly toward kidney failure and can take the time to fully educate and coach the patient through making the best renal replacement therapy option for them, whether this is home dialysis, in-center dialysis, preemptive transplant or conservative care."

MARKETPLACE

There are various vendors of predictive analytics technology on the market today. Some of these vendors include Alteryx, Anodot, Domo, Gainsight, IBM, Infer, Microsoft, Qrvey, RapidMiner, SAP, SAS Institute, Sisense and Strive Health.

MEETING THE CHALLENGE

"Strive Health's CareMultiplier platform, powered by proprietary machine learning algorithms, makes sense of massive amounts of data, cuts through the noiseand allows our clinicians to focus on doing what only they can do, deliver high-touch patient care," Dredge explained.

"Our clinical teams use predictive analytics in their day-to-day care," he continued. "Each patient receives an overall risk score that serves as a starting point for treatment and flows through our clinical care systems. As we engage our members, our team then uses focused initiatives developed through the analytics to be more proactive in their care."

As an example, SSM Health has a patient cohort called Planned Starts. Strive's technology has identified them as progressing toward dialysis in the next six to 12 months. These analytics allow clinicians to deploy focused interventions and care plans to help prevent these patients from "crashing" into dialysis.

RESULTS

"Strive Health brings economies of scale, regionalization and nationalization to a fragmented kidney care process," Dredge reported. "The program was launched in June 2020, during the COVID-19 pandemic. While the pandemic impacted most in the country, the first four months of data are promising, showing a more than 20% reduction in acute utilization for both CKD and ESRD populations and a more than 25% reduction in emergency department utilization for both CKD and ESRD populations."

Several patients have benefited from this approach, including one female patient who was predicted to have a 57% chance of kidney failure within two years. After more than a year of "watch and wait," the patient avoided a crash into dialysis through a high-touch care team coordinating between her nephrologist and primary care physician. They addressed her concerns and engaged her in appropriate treatment.

"Separately, a 36-year-old patient had acquired 16 hospital stays in two years with frequent readmissions and declining health," Dredge recalled. "This patient has since had only one emergency department visit and zero readmissions, reducing inpatient days by about 14 times her previous usage."

ADVICE FOR OTHERS

"As health systems move into population health and value-based contracts, analytics are needed to identify patient populations and follow them through their care journey," Dredge advised. "When selecting a partner, ensure there is alignment on goals and metrics.

"Understand what the healthcare organization should own versus accomplish with a partner," he continued. "Controlling all aspects of care through internal resources can stifle innovation. SSM Health's transformation team recognizes that a partner delivering an external, dedicated focus with tight integration and collaboration can speed innovation and raise all involved together for a better experience."

This leads to a virtuous cycle of innovation where the more successful one is at making progress, the faster they can go, he added.

"SSM Health turned to a partner so it could dedicate its efforts to what the health system does well, which is providing quality care to its communities," he concluded. "The partnership applied a dedicated focus to informing care that is innovating kidney care."

Twitter:@SiwickiHealthITEmail the writer:bsiwicki@himss.orgHealthcare IT News is a HIMSS Media publication.

Link:
SSM Health innovates kidney care with predictive analytics and machine learning - Healthcare IT News

Check out this topical video from Predictive Analytics World founder Eric Siegel:

A computer can keep you in jail, or deny you a job, a loan, insurance coverage, or housing and yet you cannot face your accuser. The predictive models generated by machine learning to drive these weighty decisions are generally kept locked up as a secret, unavailable for audit, inspection, or interrogation. The video above covers explainable machine learning and the loudly-advocated machine learning standards transparency and the right to explanation. Eric discusses why these standards generally are not met and overviews the policy hurdles and technical challenges that are holding us back.

About the Author

Eric Siegel, Ph.D.,is a leading consultant and former Columbia University professor who makes machine learning understandable and captivating. He is the founder of thePredictive Analytics WorldandDeep Learning Worldconference series, which have served more than 17,000 attendees since 2009, the instructor of the acclaimed online courseMachine Learning Leadership and Practice End-to-End Mastery, a popular speaker whos been commissioned formore than 110 keynote addresses, and executive editor ofThe Machine Learning Times. He authored the bestsellingPredictive Analytics: The Power to Predict Who Will Click, Buy, Lie, or Die, which has been used in courses at more than 35 universities, and he won teaching awards when he was a professor at Columbia University, where he sangeducational songsto his students. Eric also publishesop-eds on analytics and social justice. Follow him at@predictanalytic.

Read the original post:
Explainable Machine Learning, Model Transparency, and the Right to Explanation Machine Learning Times - The Predictive Analytics Times

In full transparency, the following is a press release submitted to SOURCE media through its business wire service.

***

NATICK MathWorksannounced it has been recognized for the second consecutive year as a Leader in the Gartner Magic Quadrant for Data Science and Machine Learning Platforms. MathWorks was positioned as a 2021 Leader based on the Gartner evaluation of the companys completeness of vision and ability to execute.

We believe that being recognized as a Leader for the second straight year further validates MathWorks ability to provide a comprehensive platform for solving AI challenges, said Jim Tung, MathWorks Fellow. Building on our deep experience and relentless focus on providing software and services to accelerate technical innovation, MathWorks empowers engineers and scientists to build better AI datasets, integrate community AI models, rapidly iterate and continuously test AI models in a system-wide context.

With the MATLAB technical computing platform, organizations can:

***

MathWorks is the leading developer of mathematical computing software. MATLAB, the language of engineers and scientists, is a programming environment for algorithm development, data analysis, visualization, and numeric computation. Simulink is a block diagram environment for simulation and Model-Based Design of multidomain and embedded engineering systems. Engineers and scientists worldwide rely on these product families to accelerate the pace of discovery, innovation, and development in automotive, aerospace, electronics, financial services, biotech-pharmaceutical, and other industries. MATLAB and Simulink are also fundamental teaching and research tools in the worlds universities and learning institutions. Founded in 1984, MathWorks employs more than 4500 people in 16 countries, with headquarters in Natick.

See more here:
MathWorks Named a Leader in the 2021 Gartner Magic Quadrant for Data Science and Machine Learning Platforms - framinghamsource.com

In terms of turnover, growth rate, sales, market share and scale, the Machine Learning Chip Market is rising at a fast pace and contributes significantly to the global economy. It is a detailed research paper that provides readers with useful information to understand the fundamentals of the Machine Learning Chip market. Business strategies, consumer needs, dominant market players, and a futuristic view of the market are defined in the study.

Sample Copy of This Report: https://www.quincemarketinsights.com/request-sample-64000?utm_source=Pooja/neighborwebsj

Machine Learning Chip

Effect of COVID-19

In order to reflect the most recent economic scenario and market size regarding the ongoing COVID-19 pandemic, the report was revised. In a post-COVID setting, the study looks at the development forecast as well as current and futuristic earnings projections. The study also covers the evolving patterns and dynamics of the industry as a result of the pandemic and offers an accurate overview of the effect of the epidemic on the entire market.

Main Market Study Features

Overview of Market Study:

This study offers a systematic and detailed look at the different companies working to gain a high market share in the global market for Machine Learning Chips. The report gives data for the top and fastest growing segments. A balanced combination of primary and secondary research methodologies is used to prepare this research study. According to main parameters, markets are classified, and to this end, a section dedicated to the business profile is included in the study.

Business Research Insights:

A succinct overview of the growth factors affecting the current Machine Learning Chip market scenario across different regions is put together in the study. In order to present an ensemble forecast, significant information relating to the scale, share, application, and statistics of the industry analysis is summarized in the study.

Market Segmentation

A fundamental overview of concepts, the trade lifecycle, applications and function of the trade chain are included in the market review. These variables can help leading players understand the complexity of the market, meet the needs of their customers, and deliver unique characteristics. The report provides geographic perspectives and data related to revenue and market share.

The market analysis involves the rate of expansion and related information over the forecast period. The report helps understand the revenue generated by the industry over the entire forecast period. It consists of information relating to the dynamics of the market, such as growth opportunities, the challenges involved during this vertical phase, and market factors.

Market Segmentation: By Chip Type (GPU, ASIC, FPGA, CPU, Others), By Technology (System-on-chip, System-in-package, Multi-chip module, Others), By Industry Vertical (Media & Advertising, BFSI, IT & Telecom, Retail, Healthcare, Automotive & Transportation, Others)

Get ToC for the overview of the premium report @ https://www.quincemarketinsights.com/request-toc-64000?utm_source=Pooja/neighborwebsj

Regional assessment:

In order to better understand the supply and demand ratio of this sector, the study also incorporates a region-wise segmentation analysis. A regional market analysis of North America, Europe, Asia Pacific, Middle East & Africa, And South America was carried out based on area segmentation. The present and future market scenario and the industry trends that affect the growth of the segments have also been analyzed in this exclusive analysis of the Machine Learning Chip market report.

Market Competitive Growth:

The assessment offers a competitive advantage and an appreciation of theMachine Learning Chip marketplace and the strategies for achieving significant market size in the global market. AMD (Advanced Micro Devices), Google Inc., Intel Corporation, NVIDIA, Baidu, Bitmain Technologies, Qualcomm, Amazon, Xilinx, Samsung. are some of the main players in the industry, which are analyzed in this report.

Make an Enquiry for purchasing this Report @ https://www.quincemarketinsights.com/enquiry-before-buying-64000?utm_source=Pooja/neighborwebsj

Highlights of the Market

This Machine Learning Chip market report will help you recognize your needs, identify problem areas, find better opportunities, and help all the key leadership processes in your company. To remain one step ahead and minimize losses, you can ensure the efficiency of your public relations activities and track consumer objections. Finally, the researchers shed light on various ways to explore the strengths, vulnerabilities, opportunities, and risks that impact the development of the Global Machine Learning Chip market. In this research study, the viability of the new report is also evaluated.

What insights will readers obtain from the report on the Machine Learning Chip Market?

Overview of Machine Learning Chip Market and forces propelling and restraining market growth Up-to-date analyses of market trends and technological improvements Pin-point analyses of Machine Learning Chip Market along with competition dynamics to offer competitive edge to market players Key strategies implemented by major competitors An array of graphics and SWOT analysis of major industry segments Detailed analyses of Machine Learning Chip Market trends A definite industrial growth map with an impact-analysis Offers a knowledgeable data base of the competitive landscape and key product segments This report is customized by segment, by sub-segment, by region/country, along with a product specific competitive analysis to meet your specific requirements.

ABOUT US:

QMI has the most comprehensive collection of market research products and services available on the web. We deliver reports from virtually all major publications and refresh our list regularly to provide you with immediate online access to the worlds most extensive and up-to-date archive of professional insights into global markets, companies, goods, and patterns.

Contact:

Quince Market Insights

Office No- A109

Pune, Maharashtra 411028

Phone: APAC +91 706 672 4848 / US +1 208 405 2835 / UK +44 1444 39 0986

Email: [emailprotected]

Web: https://www.quincemarketinsights.com

https://neighborwebsj.com/

Read the original post:
Machine Learning Chip Market How the Market has witnessed Substantial Growth in recent years? NeighborWebSJ - NeighborWebSJ

This market intelligence study is an extensively detailed assessment of the Data Science and Machine-Learning Platforms market and aids the client to navigate the global market landscape while upping its revenue generation potential and keeping up with the upward growth trends. This is a complete evaluation of all the market dynamics and aspects related to the Data Science and Machine-Learning Platforms market.

Get Sample PDF Brochure@ https://www.reportsintellect.com/sample-request/1039572

Leading Players Profiled in the Report: SAS, Databricks, RapidMiner, Alteryx, Dataiku, IBM, MathWorks, Microsoft, KNIME, TIBCO Software, Domino Data Lab, Rapid Insight, H20.ai, Angoss, Google, Anaconda, Lexalytics, SAP.

The market study descriptively analyzes various market dynamics and aspects that are crucial to post good growth in terms of revenue as well as overall market propulsion in the global Data Science and Machine-Learning Platforms market landscape. The market study equips the client with a detailed account of the market including the complete history along with economic forecast as well.

The Data Science and Machine-Learning Platforms report highlights the Types as follows:

Open Source Data Integration Tools, Cloud-based Data Integration Tools, etc.

The Data Science and Machine-Learning Platforms report highlights the Applications as follows:

Small-Sized Enterprises, Medium-Sized Enterprise, Large Enterprises, etc.

The report studies the following Geographical Regions:

North America (United States, Canada and Mexico)Europe (Germany, France, UK, Russia and Italy)Asia-Pacific (China, Japan, Korea, India, Southeast Asia and Australia)South America (Brazil, Argentina)MENA (Saudi Arabia, UAE, Turkey and South Africa)

Get the Discounted report @ https://www.reportsintellect.com/discount-request/1039572

Market Rivalry

This intelligence study details company profiles, product picture and specification, capacity, production, price, cost, revenue and contact information regarding the competitive landscaper of the Data Science and Machine-Learning Platforms market. The report also evaluates the market through geographical regions to provide you with more accurate data regarding each segment in the respective region.

Major factors covered in the report:

About Us:

Reports Intellect is your one-stop solution for everything related to market research and market intelligence. We understand the importance of market intelligence and its need in todays competitive world.

Our professional team works hard to fetch the most authentic research reports backed with impeccable data figures which guarantee outstanding results every time for you.So whether it is the latest report from the researchers or a custom requirement, our team is here to help you in the best possible way.

Contact Us:

sales@reportsintellect.comPhone No: + 1-706-996-2486US Address:225 Peachtree Street NE,Suite 400,`Atlanta, GA 30303

Here is the original post:
New Report on the Data Science and Machine-Learning Platforms Market (Covid-19 Updated) explores the latest trends and a details a forecast till 2027...

Social media apps are proved to be helpful to people in a lot of ways, but like everything it does have its merit and demerits as well. News and content that can be triggering to some people is also seen there and that is why it is so important to make sure that such content gets put away immediately. It is up to the makers of such apps to come up with ways to help eliminate such content. Such include content on topics like: drugs, graphic violence, adult content, medical misinformation, self-harm activities and a lot more.

One of the apps that took on the initiative to help eliminate such content was none other than Pinterest. In 2019, the image sharing social media app Pinterest launched AI (Artificial Intelligence) and machine learning technology to help with this mission on the apps platform. Such technologies help in eliminating such content by detecting content related to those categories and reporting it back. It is reported by Pinterest Engineers that due to AI and machine learning technologies there has been a decline of about 52% in unsafe content since 2019 and content encouraging self-harm have reduced to about 80% since April of 2019.

Pinterest uses a system that already is trained on millions of human reviewed Pins. The Trust and Safety operations team at Pinterest which overlook at the content detected by the machine and then assign the next actions categorize these content. How the technology helps with detecting and removing harmful content is that the company has developed a Pin model trained that is a model friendly replica of Pins based on the apps keywords and images; this then detects which content might be under violation and which is good to go by generating a score for that content with the help of another model. The engineers who are the brains behind such technologies have a hard time coming up with simpler and smaller models of such technologies that can be used in multi model inputs.

Pinterest says that technologies like such that have a complex system developed in them help with increasing more of the positive content on the platform and such technologies have helped greatly with the app becoming a much lesser toxic platform than others as also seen in the reports supporting these statements.

It is expected that other social media apps also take on initiatives like these and work to develop such technologies to help make the social media platforms a much safer, healthier and toxic free environment as possible.

Continue reading here:
Pinterest releases the details on how it's AI and machine learning technology helps the app against eliminating harmful and negative content from the...

At the recent tinyML Summit 2021, Raspberry Pi co-founder Eben Upton teased the future of 'Pi Silicon' and it looks like machine learning could see a massive improvement thanks to Raspberry Pi's news in-house chip development team.

It is safe to say that the Raspberry Pi Pico and its RP2040 SoC have been popular. The Pico has only been on the market for a few weeks, but already has sold 250,000 units with 750,000 on back order. There is a need for more boards powered by the RP2040 and partners such as Adafruit, Pimoroni, Adafruit and Sparkfun are releasing their own hardware, many with features not found on the Pico.

Raspberry Pi's in house application specific integrated circuit (ASIC) team are working on the next iteration, and seems to be focused on lightweight accelerators for ultra low power machine learning applications.

During Upton's talk at 40 minutes the slide changes and we see "Future Directions" a slide that shows three current generation 'Pi Silicon' boards, two of which are from board partners, SparkFun's MicroMod RP2040 and Arduino's Nano RP2040 Connect. The third is from ArduCam and they are working on the ArduCam Pico4ML which incorporates machine learning, camera, microphone and screen into a the Pico package.

The last bullet point hints at what the future silicon could be. It may come in the form of lightweight accelerators possibly 4-8 multiply-accumulates (MACs) per clock cycle. In Upton's talk he says that it is "overwhelmingly likely that there will be some other piece of silicon from Raspberry Pi".

Want to learn more about the Raspberry Pi Pico? We have just the page for you.

The rest is here:
Next Raspberry Pi CPU Will Have Machine Learning Built In - Tom's Hardware