Category Archives: Ai

Earlier this month we told you about AI Dungeon, an AI-powered text adventure with near infinite possibilities. You can type what you want into the game, and the AI will generate a response on the fly, creating a freewheeling experience that encourages cooperation and imagination. Now, AI Dungeon is available on iOS and Android as well, making it much easier to explore fantasy and sci-fi realms with an AI game master.

We tested the iOS version briefly, and although there were a few game-breaking errors, its generally as easy to use as youd want. Responses to each input still take a few seconds to process, which rather slows the experience, but its definitely quicker than the web version was, and each interaction with the AI is as surprising (and frequently delightful) as before.

There are also helpful tips for newcomers, reminding you to start each of your text commands with a verb, or use quotation marks to indicate when someone is speaking.

AI Dungeons creator Nick Walton told The Verge that hes actually quit his job to go full time on the game, and has started a company around it with a few other people. Hes also running a Patreon to support development, and has attracted a healthy $10k a month.

Walton says hes still surprised by the enthusiastic reaction to AI Dungeon, which is lauded by users as they find new ways to interact with the AI. I thought people would enjoy it but Ive been really blown away by how much, he says, adding that reading users play-throughs never gets old, as each person works with or against the AI in a different way.

There was one where someone told the NPCs in the game that they were in a game, says Walton. The NPCs got really depressed and sad which made them feel surprisingly lifelike.

The first step in development is to make AI Dungeon faster and more stable, he says. After that he wants to add extra features like a multiplayer mode and a family-friendly version. (Right now, AI Dungeon works just fine as a fantasy porn role-playing game. Do with that information what you will.) And after that, Walton says he and his new colleagues have much bigger plans in the AI game world, but wont say what right now. Exploring the dungeon is only the beginning.

View post:

Infinite text adventure AI Dungeon is now available on iOS and Android - The Verge

French artificial intelligence (AI) startup Iktos has announced a collaboration with global skin-health focused pharmaceutical company Almirall.

The partnership will see Almirall use Iktos generative modelling technology to optimise advanced compounds, bringing speed and efficiency to the drug discovery process.

WHY IT MATTERS

The pharma industry is increasingly turning to AI to accelerate the traditionally long and expensive process of discovering and developing new drugs.

Almirallwill use Iktos AI technology to design virtual novel molecules that have the desirable characteristics of a novel drug candidate. This will enable rapid, iterative identification of molecules, which simultaneously validate multiple bioactive attributes and drug-like criteria for clinical testing.

THE LARGER CONTEXT

Incorporated in October 2016, Iktos specialises in the development of AI solutions applied to medicinal chemistry and new drug design.

The startup recently announced several collaborations with biopharmaceutical companies which will use its AI technology to accelerate the design of promising compounds. Its generative modelling SaaS software, MakyaTM, is available on the market, and Iktos intends to release its retrosynthesis SaaS platform SpayaTM as a beta version, before the end of 2019.

Meanwhile, Swiss pharma firm Novartisrecently announced it is teaming up with Microsoft on a new AI initiative focused on drug discovery and development.It will employ Microsofts AI tools to help its scientists with computational hurdles in life sciences including generative chemistry, image segmentation and analysis for smart and personalised cell and gene therapies.

ON THE RECORD

Dr Bhushan Hardas, executive vice president R&D and chief scientific officer of Almirall, said: This partnership is an example of how we intend to explore the enormous possibilities offered by technology to find new molecules and to speed up clinical development.

The health sector lags behind others in the digital world. Almirall wants to be at the forefront of innovation to develop holistic and transversal approaches. Artificial Intelligence will provide Almirall a unique opportunity to combine our proficiency with the preciseness and celerity to truly make a difference in patients' lives.

Yann Gaston-Mathe, president and CEO of Iktos, said: This new collaboration is further testimony to the leadership position that Iktos has developed in the field of AI for de novo drug design, in little more than two years of existence.

We are eager to demonstrate to our collaborators the power of Iktos technology to accelerate their research, and to get the opportunity to further improve by confronting our approach to a new use case, consistently with our strategy to prove our value in real-life projects.

Here is the original post:

French startup Iktos and pharma firm Almirall announce research collaboration in AI for new drug design - Healthcare IT News

Paige raises $45M in Series B funding led by Healthcare Venture Partners with participation from Breyer Capital, Kenan Turnacioglu, and others.

The funding will be used to accelerate commercial efforts of its AI-native digital pathology ecosystem in the U.S., Europe, Brazil, and Canada.

Paige, a NYC-based leader in computational pathology transforming the diagnosis and treatment of cancer, today announced it has closed its Series B funding round of $45 million, bringing the Companys total capital raised to over $70 million. Healthcare Venture Partners brought the largest contribution to the round, with Breyer Capital, Kenan Turnacioglu, and other funds participating. Paige will use this new capital to drive FDA clearance of its products and expand its portfolio, delving deeper into cancer pathology, novel biomarkers, and prognostic capabilities. Additionally, the Company will accelerate commercial efforts in the U.S. and expansion in Europe, Brazil, and Canada.

Impact of Pathology on Cancer Diagnosis

Pathology is the cornerstone of cancer diagnoses. The field is on the cusp of a revolution towards digital, augmented clinical analysis. Paige aims to leverage cutting-edge AI and a vast, proprietary dataset to provide powerful new insights to pathologists, researchers, and pharmaceutical development teams.

Transforming the Diagnosis and Treatment of Cancer

Founded in 2018, Paiges mission is to revolutionize the diagnosis and treatment of cancer by providing pathologists, clinicians and researchers with insights drawn from decades of data diagnosed by world experts in cancer care. Spun out ofMemorial Sloan Kettering, Paigebuilds powerful, clinical-grade computational technologies to transform the diagnosis, treatment and biomarker discovery for cancer. With AI positioned to open a new future of pathology, Paige has created an AI-native digital pathology ecosystem that enables the Pathologist to achieve higher quality, faster throughput, and lower cost diagnosis and treatment recommendations. Additionally, Paige accelerates new biomarker discovery and is built to generate new insights into pathways and drug efficacy.

Medical AI at an Unprecedented Scale

The company plans to deliver the powerful technology via partnerships, such as the recently announced Philips deal and Paiges own AI-native platform. Paige has a comprehensive license with MSK and exclusive rights to its library of 25 million pathology slides one of the largest tumor pathology archives. Paige plans to build on to MSKs efforts and digitize millions of archived slides. This digital treasure, along with anonymized clinical data, allows them to train models at scale, and uncover new connections between Pathology, genomics, treatment response, and patient outcomes. In addition, Paige offers custom solutions for drug development teams: from pre-clinical modules to automated pathology analysis for clinical trials, and biomarker development, we are creating new possibilities to expedite and better inform teams bringing new therapeutics to the market.

The funding comes on the heels of a milestone year: Paige achieved the first FDA breakthrough designation for AI technology in Pathology and Oncology and later received the first CE mark in the space, added Thomas Fuchs, Founder of Paige and a researcher at Memorial Sloan Kettering (MSK). The Company also grew its digital slide archive to more than 1.2M images and is developing systems to combine digital slides with genomic, drug response and outcome information to create powerful new diagnostic solutions.

See original here:

Paige Raises $45M to Expand AI-Native Digital Pathology Ecosystem to Accelerate Biomarker Discovery - HIT Consultant

Oto, a startup spun off from research at SRI International to help customer service operations understand voice intonation, announced a $5.3 million seed round today.

Participants in the round included Firstminute Capital, Fusion Fund, Interlace Ventures, SAP.iO and SRI International . The total includes a previous $1 million seed round, according to the company.

Teo Borschberg, co-founder and CEO at Oto, says the company launched out of SRI International, the same company where Apples Siri technology was originally developed. It has been developing intonation data, based originally on SRI research, to help customer service operations respond better to callers emotions. The goal is to use this area of artificial intelligence to improve interactions between customer service reps (CSRs) and customers in real time.

As part of the research phase, the company compiled a database of 100,000 utterances from 3,000 speakers, culled from two million sales conversations. From this data, it has built a couple of tools to help customer service operations automate intonation understanding.

The first is a live coaching tool. Its difficult to have management monitor every call, so only a small percentage gets monitored. With Oto, CSRs can get real-time coaching on every call to raise their energy or to calm a frustrated customer before a problem escalates. In real time, were able to guide the agents on how they sound, how energetic they are, and we can nudge and push them to be more energetic, Borschberg explained.

He says this has three main advantages: more engaged agents, higher sales conversion rates and better satisfaction scores and cost reduction.

The other product measures the quality of a customer experience and gives a score at the end of each call to help the CSR (and their managers) understand how well they did, simply based on intonation. It displays the score in a dashboard. Were building a universal understanding of satisfaction from intonation, where we can learn acoustic signatures that are positive, neutral, negative, Borschberg said.

He sees a huge market opportunity here, pointing to Qualtrics, which sold to SAP last year for $8 billion. He believes that surveying people is just a part of the story. You can build a better customer experience when you understand intonation of just how well that experience is going, and you put it on a scale so that it makes it easy to understand just how well or how poorly you are doing.

The company has 20 employees today, with offices in New York, Zurich and Lisbon. It has seven customers working with the product so far, but it is still early days.

Go here to see the original:

Oto snags $5.3M seed to use AI to understand voice intonation - TechCrunch

The US government must spend $25 billion on artificial intelligence research by 2025, stem the loss of foreign AI talent, and find new ways to prevent critical AI technology from being stolen and exported, according to a policy report issued Tuesday. Otherwise it risks falling behind China and losing its standing on the world stage.

The report, from the Center for New American Security (CNAS), is the latest to highlight the importance of AI to the future of the US. It argues that the technology will define economic, military, and geopolitical power in coming decades.

Advanced technologies, including AI, 5G wireless services, and quantum computing, are already at the center of an emerging technological cold war between the US and China. The Trump administration has declared AI a national priority, and it has enacted policies, such as technology export controls, designed to limit Chinas progress in AI and related areas.

The CNAS report calls for a broader national AI strategy and a level of commitment reminiscent of the Apollo program. If the United States wants to continue to be the world leader, not just in technology but in political power and being able to promote democracy and human rights, that calls for this type of effort, says Martijn Rasser, a senior fellow at CNAS and the lead author of the report.

Rasser and his coauthors believe AI will be as pervasive and transformative as software itself has been. This means it will be of critical importance to economic success as well as military might and global influence. Rasser argues that $25 billion over five years is achievable, and notes that it would constitute less than 19 percent of total federal R&D in the 2020 budget.

We're back in an era of great power competition, and technology is that the center, Rasser says. And the nation that leads, not just artificial intelligence but technology across the board, will truly dominate the 21st century.

Over the past three years, the White Houses Office of Science and Technology Policy, which shapes the administration's technology strategy, has highlighted the importance of AI and called for federal funds to be redirected towards its development. In its 2020 budget plan, the administration has proposed $5 billion in funding for AI research and development. But officials have generally maintained that the private sector should also play a primary role in investing and developing AI.

Rasser says this is a mistake because private companies do not invest in the kind of fundamental research that serves as a foundation for big technological advances. Investments that the US government made in the 50s, 60s, and 70s propagated the technologies that the American economy, and the global economy, are built on, he says.

Artificial intelligence could prove particularly important to Americas military standing. Last month, a report to Congress by the National Security Commission on Artificial Intelligence, declared that AI will be critical to US national security. Collaboration between the government and the US tech sector would be vital, it advised.

Both the Russians and the Chinese have concluded that the way to leapfrog the US is with AI, says Bob Work, a distinguished senior fellow at CNAS who served as deputy secretary of defense under Presidents Obama and Trump. Work says the US needs to convince the public and that it doesnt intend to develop lethal autonomous weapons, only technology that would counter the work Russia and China are doing.

In addition to calling for new funding, the CNAS report argues that a different attitude towards international talent is needed. It recommends that the US attract and retain more foreign scientists by raising the number of H1-B visas and removing the cap for people with advanced degrees. You want these people to live, work, and stay in the United States, Rasser says. The report suggests early vetting of applications at foriegn embassies to identify potential security risks.

The administrations current immigration strategy arguably undermines US competitiveness. Trumps travel ban, which prevents anyone from Iran, Iraq, Somalia, Libya, Somalia, Syria, Yemen, Chad, North Korea, and Venezuela from coming to the United States, has contributed to a culture unwelcoming to foreign scientists, prompting some to study and work elsewhere. A more rigorous visa vetting process has also reduced the number of Chinese scientists that are able to enter the country.

Continue reading here:

Washington Must Bet Big on AI or Lose Its Global Clout - WIRED

I spent many years working with start-ups and large corporations, and invariably they all spent a large proportion of their budgets getting their message across to consumers and other businesses.

Yet while there is no doubt that external communications are important, I found that the most successful companies were, in fact, the ones which dedicated just as much time, effort, and resources (if not more) to get their internal communications right.

Because its all well and good projecting the right image to the outside world, but if your messaging is not consistent internally, in other words, if your mission is not clear to your own team, you have very little chance of it ringing true to those outside it.

One of my favorite quotes from Richard Branson really sums it up: Clients do not come first. Employees come first. If you take care of your employees, they will take care of the clients.

Building a cohesive and inclusive corporate culture is challenging for any company, but those challenges are exacerbated even further in an age where many employees work remotely, often in different time zones.

It might seem counter-intuitive, but where it comes to fostering better communication channels amongst humans, machines can be our most powerful allies. With technologies such as Artificial Intelligence in particular, Natural Language Processing (NLP) it is possible to enable better communication without having to over-stretch resources by dedicating people to maintaining those channels manually.

NLP has advanced significantly in recent years, becoming an increasingly sophisticated tool in discerning syntax (the arrangement of words in a logical manner) and semantics (referring to the contextual meaning of words), able to more accurately gauge human sentiment and leverage it appropriately. It can be used to bridge Top-Down Communication in Organizations and foster a better employee-employer relationship by ensuring that the employees receive all the relevant messages that they require, whether it is through regular email messaging or through highlighting their accomplishments.

It is a well-known fact that many companies fail to pass down relevant messages to their employees making them feel unwanted and disengaged, said Gaurav Bhattacharya and Saumya Bhatnagar Co-founders of Involvesoft, a platform that enables companies to improve employee engagement through the use of NLP. He stresses that nowadays it is possible for chatbots to be highly effective in establishing a meaningful connection with an employee, understanding their strengths and weaknesses and helping them where needed.

Involvesoft provides its users with an Instagram-like feed that allows employees to read the latest news and announcements, spotlight stories, and take surveys and communicate with members of their community, which in turn improves top-down communication as the messages reach all the employees through the platform.

Whereas an employee might be reluctant to ask a colleague or supervisor about something for which they feel they might get judged, a chatbot could help bridge that gap, says Bhattacharya, adding that many times an employee will refrain from asking an important question because they are worried about how friends or colleagues might judge them, and this has a significantly negative impact on productivity levels, not least because failing to ask a question might mean a mistake is repeated without that person even realizing it.

As the CEO of the company, Bhattacharya very much practices what he preaches, since his own company also leverages NLP to make employees happy and successful by helping others. The Involvesoft platform helps companies and their employees to give back to the community by creating personalized giving and volunteering opportunities.

Corporate Social Responsibility (CSR) is a hugely successful tool in building a cohesive company culture, often much more so than perks and personal rewards. In fact, Bhattacharya explains that such initiatives have been shown to improve company culture and team collaboration by up to 81% reducing turnover by as much as 30%.

NLP can thus create a self-learning virtuous circle, whereby every conversation, every interaction, and every good deed helps to evolve both humans and algorithms into a better platform. The information and insights gathered through this constant feedback loop are constantly deployed in optimization across the board within an organization: From user experience to consumer outreach, advertising and HR.

With the right NPL tools, interactions can, therefore, be turned into meaningful engagement at scale, and provide a veritable goldmine of information for human resources and marketing teams alike, improving not only the work environment but also the product. Its a win-win

This post is part of our contributor series. The views expressed are the author's own and not necessarily shared by TNW.

Published December 17, 2019 16:23 UTC

Original post:

Communication in the Age of AI - The Next Web

WSU Professor Diane Cook is one of the universitys researchers working to create a test for measuring AI as part of a project funded by DARPA. (WSU Photo)

Artificial intelligence can do a lot of impressive things, like find snow leopards among Himalayan grasses captured by remote cameras, maneuver self-driving cars through traffic, and defeat world-class opponents in the game Go.

But are these systems actually intelligent, as humans perceive the concept?

Researchers at Washington State University in Pullman are developing an IQ test to challenge AI systems to see what they really know.

We have AI systems out there that are getting really good at a variety of tasks, said WSU regents professor Diane Cook. But those feats tend to be narrow within each system. Is it really intelligent because its just learned to do that one task?

Cook and Larry Holder, both of whom are professors in WSUs School of Electrical Engineering and Computer Science, recently received a $1 million grant that will run up to five years to tackle the question. The money comes from the U.S. militarys Defense Advanced Research Projects Agency, or DARPA.

The funding began a month ago, and the researchers are starting with basic questions about the scope of intelligence, which could include recognizing images, understanding and generating natural language, reasoning, and using planning in problem solving. The scientists want to use rigorous measures, such as the ability to respond to novel experiences and transfer knowledge to different situations. They also want to test for bias in a machines knowledge; bias can lead to racial, gender and other forms of discrimination, depending on the algorithms application.

Its a difficult task to define and measure intelligence. Just look at how hard it has been to come up with effective standardized tests to measure the full range of smarts for students or job applicants.

If youre trying to see if your machine has general intelligence, you have to define what you mean by general intelligence and make sure your test is really testing that, said Melanie Mitchell, a Portland State University professor in the Department of Computer Science who is not part of the DARPA project.

One of the challenges in the field is the way in which machines learn. Mitchell gave an example of a student in her lab who was teaching a program to recognize photos that contain animals. It appeared to be learning the skill until the researchers realized that it wasnt the image of the creature that the algorithm was keying into, but rather the background blurriness. It turned out that the animals were typically in focus while the background was not, while landscape-only scenes had crisp backgrounds.

A lot of misunderstanding is that the machine learned to do a certain thing like play Go or recognizing objects, so we assume it learned it in the same way we do, Mitchell said. Were surprised when it didnt learn in the way we do, and it cant transfer its knowledge.

The WSU project is part of DARPAs Science of Artificial Intelligence and Learning for Open-world Novelty (SAIL-ON) program.

For AI systems to effectively partner with humans across a spectrum of military applications, intelligent machines need to graduate from closed-world problem solving within confined boundaries to open-world challenges characterized by fluid and novel situations, says the SAIL-ON website. The programs goal is to create and test high-performing AI systems to meet the militarys needs.

There are other organizations working to expand and understand AI abilities. In the Northwest that includes Seattles Allen Institute for Artificial Intelligence (AI2) and the AI group at the University of Washingtons Paul G. Allen School of Computer Science and Engineering. In September, AI2 announced that it had built an AI program called Aristo that is smart enough to pass an eighth-grade, multiple-choice science test.

WSUs Holder has an Artificial Intelligence Quotient or AIQ website with some initial tests for AI developers to quiz their systems. The site is a publicly available tool that will also provide data to the researchers.

We are focused on testing and improving systems that can be more general-purpose, like a robot assistant that can help you with many of your day-to-day tasks, Holder said in a prepared release.

The WSU scientists aim to create a test that will grade AI technology according the difficulty of the problems it can solve. Initial plans for tests include video games, answering multiple choice problems and solving a Rubiks cube.

Its an opportunity, said Cook, to get back the grassroots and say what AI is.

More:

Whats the best way to measure the smarts of AI systems? Researchers are developing an IQ test - GeekWire

There are about 7.3 million open jobs in the U.S., according to the most recent Job Openings and Labor Turnover Survey from the Bureau of Labor Statistics. And for many job seekers vying for these openings, the likelihood they'll submit their application to an artificial intelligence-powered hiring system is growing.

A 2017 Deloitte report found 33% of employers already use some form of AI in the hiring process to save time and reduce human bias. These algorithms scan applications for specific words and phrases around work history, responsibilities, skills and accomplishments to identify candidates who match well with the job description.

These assessments may also aim to predict a candidate's future success by matching their abilities and accomplishments to those held by a company's top performers.

But it remains unclear how effective these programs are.

As Sue Shellenbarger reports for The Wall Street Journal, many vendors of these systems don't tell employers how their algorithms work. And employers aren't required to inform job candidates when their resumes will be reviewed by these systems.

That said, "it's sometimes possible to tell whether an employer is using an AI-driven tool by looking for a vendor's logo on the employer's career site," Shellenbarger writes. "In other cases, hovering your cursor over the 'submit' button will reveal the URL where your application is being sent."

CNBC Make It spoke with career experts about how to make sure your next application makes it past the initial robot test.

AI-powered hiring platforms are designed to identify candidates whose resumes match open job descriptions the most. These machines are nuanced, but their use still means very specific wording, repetition and prioritization of certain phrases matter.

Job seekers can make sure to highlight the right skills to get past initial screens by using tools, such as an online cloud generator, to understand what the AI system will prioritize most. Candidates can drop in the text of a job description and see which words appear most often, based on how large they appear within the word cloud.

CareerBuilder also created an AI resume builder to help candidates include skills on an application they may not have identified on their own.

Including transferable skills mentioned in the job description can also increase your resume odds. After all, executives from a recent IBM report say soft skills such as flexibility, time management, teamwork and communication are some of the most important skills in the workforce today.

"Job seekers should be cognizant of how they are positioning their professional background to put their best foot forward," Michelle Armer, chief people officer at talent acquisition company CareerBuilder, tells CNBC Make It. "Since a candidate's skill set will help set them apart from other applicants, putting these front and center on a resume will help make sure you're giving skills the attention they deserve."

It's also worth noting that AI enables employers to source candidates from the entire application system more easily, rather than limiting consideration just to people who applied to a specific role. "As a result," says TopResume career expert Amanda Augustine, "you could be contacted for a role the company believes is a good fit even if you never specifically applied for that opportunity."

When it comes to actually writing your resume, here are seven ways to make sure it looks best for the robots who will be reading it.

Use a text-based application like Microsoft Word rather than a PDF, HTML, Open Office, or Apple Pages document so buzzwords can be accurately scanned by AI programs. Augustine suggests job seekers skip images, graphics and logos, which might not be readable. Test how well bots will comprehend your resume by copying it into a plain text file, then making sure nothing gets out of order and no strange symbols pop up.

Order your work history based on most senior-level role first. Reorder the bullets underneath your job titles to mimic the order of qualifications the employer gave. Augustine says machines favor documents with a clear hierarchy to their information.

Include keywords from the job description, such as the role's day-to-day responsibilities, desired previous experience and overall purpose within the organization. Consider having a separate skills section, Augustine says, where you list any certifications, technical skills and soft skills mentioned in the job description.

Quantify performance results, Shellenbarger writes. Highlight ones that involve meeting company goals, driving revenue, leading a certain number of people or projects, being efficient with costs and so on.

Tailor each application to the description of each role you're applying for. These AI systems are generally built to weed out disqualifying resumes that don't match enough of the job description. The more closely you mirror the job description in your application, the better, Augustine says.

Don't place information in the document header or footer, even though resumes traditionally list contact information here. According to Augustine, many application systems can't read the information in this section, so crucial details may be omitted.

Network within the company to build contacts and get your resume to the hiring manager's inbox directly. "While AI helps employers narrow down the number of applicants they will move forward with for interviews," Armer says, "networking is also important."

AI hiring programs show promise at filling roles with greater efficiency, but can also perpetuate bias when they reward candidates with similar backgrounds and experiences as existing employees. Armer stresses hiring algorithms need to be built by teams of diverse individuals across race, ethnicity, gender, experience and other background factors in order to minimize bias.

This is also where getting your resume in front of a human can pay off the most.

"When you have someone on the inside advocating for you, you are often able to bypass the algorithm and have your application delivered directly to the recruiter or hiring manager, rather than getting caught up in the screening process," Augustine says.

Augustine recommends job seekers take stock of their existing network and identify those who may know someone at the companies they're interested in working at. "Look for professional organizations and events that are tied to your industry 10times.com is a great place to find events around the world for every imaginable field," she adds.

Finally, Armer recommends those starting their job hunt review and polish their social media profiles.

Like this story? Subscribe to CNBC Make It on YouTube!

Don't miss: This algorithm can predict when workers are about to quithere's how

Go here to see the original:

7 tips to get your resume past the robots reading it - CNBC

VANCOUVER, British Columbia--(BUSINESS WIRE)--Top AI conference NeurIPS 2019 was held in Vancouver from December 8-14th. Attending experts were excited about a new research direction named federated learning (FL). Professor Yoshua Bengio, A.M. Turing Award Winner, founder of the worlds top deep learning research facility Mila-Quebec Artificial Intelligence Institute and one of the "three musketeers of deep learning", said that In terms of better training neural networks, federated learning is at the forefront of research and will have important impact on business.

Currently, data silos and privacy protection are two big challenges for AI. As an encrypted distributed machine learning framework, FL can tackle both problems by allowing different parties to build models collaboratively without the need to reveal their data. The method helps to advance AI modeling while protecting data and privacy.

Chinas digital bank WeBank is a leading research facility in federated learning. At NuerIPS 2019, WeBank co-organized the FL workshop with Google, CMU, and NTU, with 400 scholars joining in the discussion.

During the WeBank AI Night event, WeBank announced two strategic partnerships with Mila and the leading cloud computing platform Tencent Cloud. The cooperation will focus on further developing federated learning, based on WeBank's real-world experiences in finance and fintech, adhering to Milas core philosophy AI for Humanity, Tencents AI for Good and WeBanks Make Banking Better for All " to create safe, inclusive AI applications.

Professor Qiang Yang, WeBanks chief AI officer, explained that large-scale AI application relies on big data, which is scattered across many different organizations. Direct data merging will violate privacy regulations. FL is a compliance method strictly following laws and regulations, and is now used in fintech, healthcare, smart city, and other industrial applications.

To reduce the use threshold of federated learning, WeBank launched the world's first industry FL open-source framework Federated AI Technology Enabler (FATE) in February 2019. This grants a ready-to-use FL framework tool to any companies wishing to work together. Partner Tencent Cloud and companies including Huawei, JD.com and other tech giants have all joined the ecosystem. The company is also leading international IEEE standards on the technology.

Founded in 2014, WeBank is the worlds leading digital bank operating solely online, now serving over 170 million individual customers and over 500,000 small and micro-sized enterprises.

View original post here:

NuerIPS 2019: China's WeBank, Mila, and Tencent Partner on AI Federated Learning to Protect Data Privacy - Business Wire

AI has a black box problem. We cram data in one side of a machine learning system and we get results out the other, but were often unsure what happens in the middle. Researchers and developers nearly had the issue licked, with explainable algorithms and transparent AI trending over the past few years. Then came the lawyers.

Black box AI isnt as complex as some experts make it out to be. Imagine you have 1,000,000 different spices and 1,000,000 different herbsand you only have a couple of hours to crack Kentucky Fried Chickens secret recipe. Youre pretty sure you have all the ingredients but youre not sure which eleven herbs and spices you should use. You dont have time to guess, and it would take billions of years or more to manually try every combination. This problem cant realistically be solved using brute force, at least not under normal kitchen paradigms.

But imagine if you had a magic chicken fryer that did all the work for you in seconds. You could pour all your ingredients into it and then give it a piece of KFC chicken to compare against. Since a chicken fryer cant taste chicken, it would rely on your taste-buds to confirm whether itd managed to recreate the Colonels chicken or not.

It spits out a drumstick, you take a bite and tell the fryer whether the piece youre eating now tastes more or less like KFCs than the last one you tried. The fryer goes back to work, tries more combinations, and keeps going until you tell it to stop once it has the recipe right.

Thats basically how black box AI works. You have no idea how the magic fryer came up with the recipe maybe it used 5 herbs and 6 spices, maybe it used 32 herbs and 0 spices but, it doesnt matter. All we care about is using AI as a way to do something humans could do, but much faster.

This is fine when were using blackbox AI to determine whether something is a hotdog or not, or when Instagram uses it to determine if youre about to post something that might be offensive. Its not fine when we cant explain why an AI sentenced a black man with no priors to more time than a white man with a criminal history for the same offense.

The answer is transparency. If there is no black box, then we can tell where things went wrong. If our AI sentences black people to longer prison terms than white people because its over-reliant on external sentencing guidance, we can point to that problem and fix it in the system.

But theres a huge downside to transparency: If the world can figure out how your AI works, it can figure out how to make it work without you. The companies making money off of black box AI especially those like Palantir, Facebook, Amazon, and Google who have managed to entrench biased AI within government systems dont want to open the black box anymore than they want their competitors to have access to their research. Transparency is expensive and, often, exposes just how unethical some companies use of AI is.

As legal expert Andrew Burt recently wrote in Harvard Business Review:

To start, companies attempting to utilize artificial intelligence need to recognize that there are costs associated with transparency. This is not, of course, to suggest that transparency isnt worth achieving, simply that it also poses downsides that need to be fully understood. These costs should be incorporated into a broader risk model that governs how to engage with explainable models and the extent to which information about the model is available to others.

The AI gold rush of the 2010s led to a Wild West situation where companies can package their AI any way they want, call it whatever they want, and sell it in the wild without regulation or oversight. Companies that have made millions or billions selling products and services related to biased, black box AI have managed to entrench themselves in the same position as the health insurance and fossil fuel industries. Their very existence is threatened by the idea that they may be regulated against doing harm to the greater good.

Simply put: No. The lawyers will make sure well never know any more about why a commercial system is biased, even if we develop fully transparent algorithms, than if these systems remain in black boxes. As Axios Kaveh Waddell recently wrote:

Companies are tightening access to their AI algorithms, invoking intellectual property protections to avoid sharing details about how their systems arrive at critical decisions.

The calculus for the AI industry is the same as the private healthcare industry in the US. Extricating biased black box AI from the world would probably put dozens of companies out of business and likely result in hundreds of billions of dollars lost. The US industrial law enforcement complex runs on black box AI were unlikely to see the government end its deals with Microsoft, Palantir, and Amazon any time soon. So long as the lawmakers are content to profit from the use of biased, black box AI, itll remain embedded in society.

And we also cant rely on businesses themselves to end the practice. Our desire to extricate black box systems simply means companies cant blame the algorithm anymore, so theyll hide their work entirely. With transparent AI, well get opaque developers. Instead of choosing not to develop dual use, or potentially dangerous AI, theyll simply lawyer up.

As Burt puts it in his Harvard Business Review article:

Indeed, this is exactly why lawyers operate under legal privilege, which gives the information they gather a protected status, incentivizing clients to fully understand their risks rather than to hide any potential wrongdoings. In cybersecurity, for example, lawyers have become so involved that its common for legal departments to manage risk assessments and even incident-response activities after a breach. The same approach should apply to AI.

When things go wrong and AI runs amok, the lawyers will be there to tell us the most company-friendly version of what happened. Most importantly, theyll protect companies from having to share how their AI systems work.

Were trading a technical black box for a legal one. Somehow, this seems even more unfair.

Read next: The super-rare Nintendo Play Station prototype is going to auction

View original post here:

Researchers were about to solve AI's black box problem, then the lawyers got involved - The Next Web