Tech News

Earlier this year, a group of researchers from Osaka University introduced an AI system based on convolutional neural network (CNN) for automatically identifying radioresistant tumor cells. In a single cancer tumor, there can be tremendous variation in the cancer cells types which can make it difficult for doctors to make accurate assessments of cell types. Further, this process can be time-consuming and can often be hampered by human error. This AI-based system can make it easy for doctors to choose the most effective treatment and also finds applications in preclinical cancer research. Explaining the utility of this system, one of the researchers, Masayasu Toratani said, “The automation and high accuracy with which this system can identify cells should be very useful for determining exactly which cells are present in a tumor or circulating in the body of cancer patients. For example, knowing whether or not radioresistant cells are present is vital when deciding whether radiotherapy would be effective, and the same approach can then be applied after treatment to see whether it has had the desired effect.” For the study, the researchers used phase-contrast images of radioresistant clones for two cell lines: mouse squamous cell carcinoma NR-S1, and human cervical carcinoma ME-180. They gathered 10,000 images of each of the parental NR-S1 and ME-180 controls as well as radioresistant clones. VGG16, a convolutional neural network for object recognition, was then trained on 8,000 images of cells. For testing the model, the researchers used another 2,000 images to check its accuracy. The model was able to give an accuracy of 96%. As per the results, it had learned the features that distinguish mouse cancer cells from human ones, and radioresistant cancer cells from radiosensitive ones. The features extracted by this trained CNN were then plotted using t-distributed stochastic neighbor embedding, and the plot showed that the images of each cell line were well clustered. In the future, the researchers will train the system on different types of cell types to make it a universal system that can automatically identify and distinguish all variants of cancer cells. To know more in detail, check out the study published by Cancer Research. REVOLVER: A machine learning approach to forecast cancer growth Google, Harvard researchers build a deep learning model to forecast earthquake aftershocks location with over 80% accuracy Stanford researchers introduce DeepSolar, a deep learning framework that mapped every solar panel in the US

With more than a year of scandals and data breach from Facebook, the company leaves no stone unturned to prove itself right and follow ethics washing. The company has also been in the radar of FTC and is expected to be fined around $5 billion because of its user data practices. Last week, Facebook argued that it didn't violate users' privacy rights because there's no expectation of privacy when using social media and the company wants to dismiss a lawsuit related to the Cambridge Analytica scandal, by arguing the same. Facebook counsel Orin Snyder said during a pretrial hearing to dismiss a lawsuit, "There is no invasion of privacy at all because there is no privacy." Facebook didn't deny that third parties accessed users' data, but the company told Vince Chhabria, US District Judge that there's no "reasonable expectation of privacy" on Facebook or any other social media site. But the argument coming from Facebook appears to be more like the company is trying to convince people that it knows how to protect their personal information. This month Sheryl Sandberg, Facebook COO said that she and Mark Zuckerberg at Facebook, will do "whatever it takes" to keep people safe on Facebook. Calls to curb Zuckerberg's control over Facebook have now taken rounds as the issues around data privacy and security continue. It seems Chhabria is making sure that at least some of the lawsuit continues, saying in an order before the hearing (PDF) that the plaintiffs should expect the court to accept their argument that private information was disclosed without express consent. Facebook releases Pythia, a deep learning framework for vision and language multimodal research Facebook tightens rules around live streaming in response to the Christchurch terror attack Facebook again, caught tracking Stack Overflow user activity and data

British and Dutch authorities have fined Uber for a total of nearly $1.2m on Tuesday over a data breach incident that occurred in 2016. The Information Commissioner's Office (ICO) from UK imposed a £385,000 fine (close to $500,000) on Uber for “failing to protect customers' personal information during a cyber attack". The said attack happened in November 2016. Additionally, the Dutch Data Protection Authority imposed their own €600,000 (close to $680,000) fine over the same incident for not reporting the data breach to the Dutch DPA within 72 hours after the discovery of the breach. For the same data breach, the US government has fined Uber $148m. Attackers obtained login credentials to access Uber’s servers and downloaded files in November 2016. These files contained records of users worldwide including passengers’ full names, phone numbers, and email addresses. Personal details of around 2.7million UK customers and 174,000 Dutch citizens were downloaded from Uber cloud servers by hackers in this breach. Steve Eckersley, the Director of Investigations at ICO, said: “This was not only a serious failure of data security on Uber’s part, but a complete disregard for the customers and drivers whose personal information was stolen. At the time, no steps were taken to inform anyone affected by the breach, or to offer help and support. That left them vulnerable.” As the attack occurred in 2016, it was not subject to the EU's GDPR that came into effect May 2018. The GDPR rules could have increased the fines for Uber. The affected customers and drivers were not told about the incident and Uber started monitoring the accounts for fraud only after an year. The attackers then demanded $100,000 to destroy the data they took which Uber paid as “bug bounty”. This is unlike a legitimate bug bounty program which is a common practice in tech industries. The attackers had malicious intent hence they downloaded the data as opposed to just pointing out the breach. Eckersley further added: “Paying the attackers and then keeping quiet about it afterwards was not, in our view, an appropriate response to the cyber attack.” In a statement, Uber representatives said “We’re pleased to close this chapter on the data incident from 2016. We’ve also made significant changes in leadership to ensure proper transparency with regulators and customers moving forward. We learn from our mistakes and continue our commitment to earn the trust of our users every day.” Uber posted a billion dollar loss this quarter. Can Uber Eats revitalize the Uber growth story? EU slaps Google with $5 billion fine for the Android antitrust case Origin DApp: A decentralized marketplace on Ethereum mainnet aims to disrupt gig economy platforms like Airbnb and Uber

According to a story by Windows Central, Microsoft is working on a Chromium based web browser. This will likely be a replacement to their current web browser on Windows 10, Microsoft Edge. Edge never took off the Edge Microsoft Edge was launched in 2015 built from scratch with EdgeHTML. Microsoft tried to get it into adoption by making Windows 10 update free for a limited time. However, the browser was not well received at the early stage itself due to a large number of issues. Since then it has not been very stable driving developers and users away from it. Due to this, Microsoft is reportedly abandoning Edge and the EdgeHTML framework for Chromium. Chromium is a rendering engine used in Google Chrome. The new browser’s name is codenamed Anaheim and will be replacing Edge as the default browser in Windows. It is not known if Edge will be renamed and if the UI will be different. But EdgeHTML will no longer be used in Windows 10’s default browser. Using the Chromium engine instead Using Chromium means that the websites on Windows 10 default browser will load as they do on Google Chrome. Default browser users will no longer have to face the loading and connectivity issues that plagued EdgeHTML based Microsoft Edge. For smartphones, nothing will change much as Edge on smartphones already use platform specific engines. Recently, 9to5Google reported that Microsoft engineers are committing code to the Chromium project. This would suggest that they are working on their own browser by using Chromium instead of EdgeHTML. The browser may likely be out next year. Public reactions A comment on hacker news reads: “You[web developers] don't test your work in Edge and because you tell all your friends and family to use Chrome instead of Edge. So stop complaining about monoculture. Many of you helped create it.” Some sarcasm thrown in another comment: “I test my app in Edge, every time a new version is released. When it inevitably fails, I shake my head in disbelief that Microsoft still hasn't paid a dev to spend a couple months fixing their IndexedDB implementation, which has been incomplete since the IE days. Can't expect a small rag-tag group like Microsoft to compete with a rich corporate behemoth like Mozilla, I guess :)” Another comment says: “How can I test Edge when Microsoft don't release it for Mac and Linux? A browser for a single OS? Talk about monoculture.” https://twitter.com/headinthebox/status/1069796773017710592 Another Tweet suggests that this move towards Chromium is about ElectronJS stronghold over app development and not about Microsoft wanting a Chrome like browser experience on its default browser: https://twitter.com/SwiftOnSecurity/status/1069776335336292352 After years of Internet Explorer being ridiculed and Edge not being the success they hoped for, it would be nice to see the Windows default browser catching up with the likes of Chrome and Firefox. Introducing Howler.js, a Javascript audio library with full cross-browser support Firefox Reality 1.0, a browser for mixed reality, is now available on Viveport, Oculus, and Daydream Microsoft becomes the world’s most valuable public company, moves ahead of Apple

Yesterday, the team at Apache NetBeans released Apache NetBeans IDE 10.0, an integrated development environment for Java. This release focuses on adding support for JDK 11, JUnit 5, PHP, JavaScript, and Groovy. What’s new in Apache NetBeans IDE 10.0? JDK 11 Support Integration with the nb-javac project is now possible. This integration adds support for JDK 11. The CORBA modules have been removed. This release comes with a support for JEP 309, Dynamic Class-File Constants. It also supports JEP 323, Local-Variable Syntax  and LVTI for Lambda parameters. PHP Support PHP 7.3 It is now possible to add trailing commas in function calls under PHP 7.3. This release comes with support for Heredoc and Nowdoc Syntaxes. PHP 7.2 This release comes with support for trailing commas in list syntax and coloring for object types for PHP 7.2. PHP 7.1 This release comes with class constant visibility, multi-catch exception handling, nullable types, support for keys in list(), coloring for new keywords (void, iterable) for PHP 7.1. JUnit 5 JUnit 5.3.1 has been added as a new Library to NetBeans, so users can easily add it to their Java projects. JUnit 5 is now the default JUnit version for Maven projects without any existing tests. This release supports JUnit 5 @Testable annotation. This version also supports a default JUnit 5 test template. OpenJDK This release automatically detects JTReg from OpenJDK configuration. Various improvements such as limiting directories that are scanned for modules have been made to the OpenJDK project. Few users have compared Apache NetBeans IDE 10.0 with Eclipse and Intellij most of them are on the opinion that this release is better than the two and it works better. Read more about this release in detail on Apache NetBeans’ official blog. Apache NetBeans 9.0 is now available with Java 9 & 10 support Apache NetBeans 9.0 RC1 released! The NetBeans Developer’s Life Cycle

The year 2018 proved to be interesting for the team at Thunderbird, a free email application, as they released the latest ESR, Thunderbird 60, which had improved security, stability, and the app’s interface. However, 2019 has some upgrades to Thunderbird’s calendar. With the new year, Thunderbird has come up with some interesting plans and is working towards bringing improvements to UI, Gmail support, notifications and much more. Let’s dive deeper into the plans and goals marked by the team at Thunderbird for this year. The roadmap towards making Thunderbird better Making Thunderbird faster than ever The team is working towards addressing technical debt, UI-slowness, and general performance issues across the application (Thunderbird). They will be focusing on methods for testing and measuring slowness and working on solutions to address the pain points. They will also be working on faster technologies in rewriting parts of Thunderbird and simultaneously will be working towards a multi-process Thunderbird. Better UI and Gmail support 2019 will mark improvements in Thunderbird’s UX/UI. The team plans to focus on integration improvements in various areas. They currently have plans for better Gmail support in mind, considering that it is one of the biggest Email providers, it would definitely make sense to work on this area. While addressing Gmail label support, Thunderbird is also ensuring that other features specific to the Gmail experience translate properly into Thunderbird. This will help Thunderbird become more native on each desktop and will make managing notifications from the app easier. The team also plans to work on encrypting email and ensuring secure communication in upcoming releases. They have plans for bringing architectural changes to support smoother operations. Better notifications (system integrated) The team is improving notifications in Thunderbird by integrating with each operating system’s built-in notification system. Earlier they worked towards unifying the implementation across platforms, but this was not completely finished and might get accomplished this year. The team might drop a lot of platform dependent implementations and unify the content production logic. Improvements to rewrite and mail filters Currently the filtering is synchronously done in C++ and might be changed to async JavaScript implementation this year. Filtering will be made contextual, which means the team will be adding the ability for pre-filter MIME processing so that filtering can work on a message representation instead of on the raw MIME. Thunderbird will be addressing the problem of not having filters on mobile and ensuring that the filter can sync up to the server. Calendar improvements In 2019, Thunderbird will remove the use of all calendar XPCOM components and will replace them with simple JavaScript classes. The calendar and tasks tabs will be self-contained and will be only using HTML. The Thunderbird UI integration will be changed so that most calendar features get visible once triggered. Improved .ics handling Thunderbird will now handle inline event display better. This year will bring improvements to invites in order to see the new event details before taking action. Users are excited about the upcoming development and have their share of suggestions. One of the users commented on HackerNews saying that he would want improved native CardDAV and CalDAV support, Native PGP and much more in the coming releases. Another user commented, “Full-text indexing for PGP mail would be nice too once it's native (Mailpile and CanaryMail helped pave the way on this I believe).” Read more about the updates expected in Thunderbird’s mailing list. Microsoft’s move towards ads on the Mail App in Windows 10 sparks privacy concerns Email and names of Amazon customers exposed due to ‘technical error’; number of affected users unknown LinkedIn used email addresses of 18M non-members to buy targeted ads on Facebook, reveals a report by DPC, Ireland

Google yesterday announced a new external advisory board to help monitor the company’s use of artificial intelligence for ways in which it may violate ethical principles it laid out last summer. The group was announced by Kent Walker, Google’s senior vice president of global affairs, and it includes experts on a wide-ranging series of subjects, including mathematics, computer science, philosophy, psychology, and even foreign policy. Following is the complete list of the advisory council appointed by Google: Alessandro Acquisti, a leading behavioral economist and privacy researcher. Bubacarr Bah, an expert in applied and computational mathematics De Kai, a leading researcher in natural language processing, music technology and machine learning Dyan Gibbens, an expert in industrial engineering and CEO of Trumbull Joanna Bryson, an expert in psychology and AI, and a longtime leader in AI ethics Kay Coles James, a public policy expert with extensive experience working at the local, state and federal levels of government Luciano Floridi, a leading philosopher and expert in digital ethics William Joseph Burns, a foreign policy expert and diplomat The group will be called the Advanced Technology External Advisory Council, and it appears Google wants it to be seen as an independent watchdog keeping an eye on how it deploys AI in the real world. It wants to focus on facial recognition technology and mitigation of built-in bias in machine learning training methods. “This group will consider some of Google’s most complex challenges that arise under our AI Principles ... providing diverse perspectives to inform our work,” Walker writes. Behind the selection of the council As for the members, the names may not be easily recognizable to those outside academia. However, the credentials of the board appear to be of the highest caliber, with resumes that include multiple presidential administration positions and stations at top-notch universities spanning University of Oxford, Hong Kong University of Science and Technology, and UC Berkeley. Having said that, the selection of the Heritage Foundation President Kay Coles James and CEO of Trumbull Dyan Gibbens received harsh criticism on Twitter. It has been noted that James, through her involvement with the conservative think tank, has espoused anti-LGBTQ rhetoric on her public Twitter profile: https://twitter.com/farbandish/status/1110624709308121088 https://twitter.com/EerkeBoiten/status/1110675556713091072 One of the members, Joanna Bryson also expressed astonishing comments on Twitter for being selected as a part of the council. Joanna states, she has no idea of what she is getting into but she will certainly do her best. https://twitter.com/luke_stark/status/1110630992979652608 Google’s history of controversies Last year, Google found itself embroiled in controversy over its participation in a US Department of Defense drone program called Project Maven. Following immense internal backlash and external criticism for putting employees to work on AI projects that may involve the taking of human life, Google decided to end its involvement in Maven following the expiration of its contract. It also put together a new set of guidelines, what CEO Sundar Pichai dubbed Google’s AI Principles, that would prohibit the company from working on any product or technology that might violate “internationally accepted norms” or “widely accepted principles of international law and human rights.” “We recognize that such powerful technology raises equally powerful questions about its use,” Pichai wrote at the time. “How AI is developed and used will have a significant impact on society for many years to come. As a leader in AI, we feel a deep responsibility to get this right.” Google effectively wants its AI research to be “socially beneficial,” and that often means not taking government contracts or working in territories or markets with notable human rights violations. Regardless, Google found itself in yet another similar controversy related to its plans to launch a search product in China, one that may involve deploying some form of artificial intelligence in a country currently trying to use that very same technology to surveil and track its citizens. Google’s pledge differs from the stances of Amazon and Microsoft, both of which have said they will continue to work the US government. Microsoft has secured a $480 million contract to provide HoloLens headsets to the Pentagon, while Amazon continues to sell its Rekognition facial recognition software to law enforcement agencies. Google also formed a “responsible innovation team” internally that Walker says has reviewed hundreds of different launches to-date, some of which have aligned with its principles while others haven’t. For example, that team helped Google make the decision not to sell facial recognition technology until there’s been more ethical and policy debate on the issue. Why critics are skeptical of this move? Rashida Richardson, director of policy research at AI Now Institute, expressed skepticism about the ambiguity of Google and other companies’ AI principles at the MIT Technology Review Conference held in San Francisco on Tuesday. For example, Google’s document leans heavily on the word “appropriate.” “Who is defining what appropriate means?” she asked. Walker said that Google's new council is meant to foster more defined discussion. He added that the company had over 300 people looking at machine learning fairness issues. "We’re doing our best to put our money where our mouth is,” Kent said. Google has previously had embarrassing technology screw-ups driven by bias in its machine learning systems, like when its photos algorithm labeled black people as gorillas. It would not be wrong to say that today’s announcement — which perhaps not coincidentally comes a day after Amazon said it would earmark $10 million with the National Science Foundation for AI fairness research, and after Microsoft executive Harry Shum said the company would add an ethics review focusing on AI issues to its standard product audit checklist — appears to be an attempt by Google to fend off broader, continued criticism of private sector AI pursuits. https://twitter.com/smunson/status/1110657292549029888 Thoughtful decisions require careful and nuanced consideration of how the AI principles … should apply, how to make tradeoffs when principles come into conflict, and how to mitigate risks for a given circumstance,” says Walker in an earlier blog post. Google and Facebook working hard to clean image after the media backlash from the Christchurch terrorist attack Google announces Stadia, a cloud-based game streaming service, at GDC 2019 Google to be the founding member of CDF (Continuous Delivery Foundation)

Lyft created its Level 5 division last year with the aim to focus on developing self-driving cars. It is now acquiring London based computer vision startup Blue Vision Labs to bring safe and reliable autonomous driving to the streets first. Self-driving cars is one of the most challenging areas of applied machine learning today. In just a year, the Level 5 division into a team of 300 engineers and researchers. This acquisition marks the first acquisition for Level 5 and also its first step into the UK self-driving space. Blue Vision uses computer vision to build large-scale robotics and augmented reality applications. It was found in 2016 by graduates from University of Oxford and Imperial College London. Today they consist of 40 skilled experts in computer vision and robotics. With the technology from Blue Vision Labs, entire city maps in 3D can be formed just with the cameras mounted on cars. These maps make the car aware of its environment with high accuracy. In a Medium post, the VP of Engineering with Lyft, Luc Vincent says: “Blue Vision Labs is the first company able to build city-scale 3D maps from cell phone acquired imagery. This is truly amazing tech.” Vincent also hinted at bigger plans for the role Blue Vision Labs will play in the growth Lyft’s self driving division. He said, “It also has applications well beyond self-driving. For example, we are keen to explore how we can leverage Blue Vision Labs’ stack to more precisely pinpoint drivers’ and riders’ locations, and create new augmented reality interfaces that make transportation simpler and better for everyone.” In-vehicle advertising is a space all tech titans serious about autonomous tech like Alphabet’s Waymo and Apple’s secretive self-driving car project are vying for. Lyft seems to understand the value of being first to market in this area with this promising acquisition. Although there is no official statement about the acquisition details, as per some sources for TechCrunch, it is “around $72 million with $30 million on top of that based on hitting certain milestones.” This acquisition will drive Lyft’s self-driving visions on the streets of UK. Self-driving cars vacate an extra seat in cars and can contribute towards reducing the effects of problems like reducing pollution, traffic etc, believes the Lyft Level 5 team. To know more details about Lyft self-driving, visit the Lyft website. nuScenes: The largest open-source dataset for self-driving vehicles by Scale and nuTonomy This self-driving car can drive in its imagination using deep reinforcement learning Tesla is building its own AI hardware for self-driving cars

Last week, the Electronic Frontier Foundation (EFF) issued a letter to support the petition for review filed by Richard Sander and the First Amendment Coalition in Sander v. State Bar of California case. The opinion issued by First District Court of Appeal in August basically changes the California Public Records Act (CPRA) that could prevent California citizens from accessing public data that state and local agencies are generating. The court ruled that in order to de-identify personal information, the State Bar of California has to create “new records” to “recode its original data into new values.” EFF has raised a question that the California Supreme Court has to address: does anonymization of public data amount to a creation of new records under the CPRA? If the court’s opinion of creating new records becomes the standard across California, it will be against the purpose of CPRA. CPRA was signed in 1968, a result of a 15 year long effort to create a general records law for California. Under CPRA, on public request, the governmental records should be shared with the public, unless there is any reason not to do so. This act enables people to understand what the government is doing and prevents government inefficiencies. This act is very important today as a vast amount of digital data is produced and consumed by governments. In a previous hearing the California Supreme Court acknowledged that sharing this data to the public will prove useful: “It seems beyond dispute that the public has a legitimate interest in whether different groups of applicants, based on race, sex or ethnicity, perform differently on the bar examination and whether any disparities in performance are the result of the admissions process or of other factors.” However, when the case proceeded to trial, the petitioners were asked to show how it was possible to de-identify this data. But, according to CPRA, when government refuses to share the records requested by the public, they should show the court that it is not possible to release data and protect private information at the same time. EFF further pointed that in another case, Exide Technologies v. California Department of Public Health, a different superior court in California has ruled the opposite way. The court ruled that the government agency must share the investigations of blood lead levels. But it should be shared in a format that serves the public interest in government transparency while at the same time protecting the privacy interests of individual lead-poisoning patients. This requires California Supreme Court to settle how agencies should handle sensitive digital information under the CPRA. With the increase in the data collected by the state from and about the public, it is important that they give access to this data in order to maintain the transparency. Read the full announcement on EFF's official website. Senator Ron Wyden’s data privacy law draft can punish tech companies that misuse user data Privacy experts urge the Senate Commerce Committee for a strong federal privacy bill “that sets a floor, not a ceiling” Is AT&T trying to twist data privacy legislation to its own favor?

Yesterday, Facebook open sourced a suite of Linux kernel components and tools. This suite includes products that can be used for resource control and utilization, workload isolation, load balancing, measuring, monitoring, and much more. Facebook has already started using these products on a massive scale throughout its infrastructure and many other organizations are also adopting them. The following are some of the products that they have open sourced: Berkeley Packet Filter (BPF) BPF is a highly-flexible Linux kernel code execution engine. It enables safe and easy modifications of kernel behaviors with custom code by allowing bytecode to run at various hook points. Currently, it is being widely used for networking, tracing and security in a number of Linux kernel subsystems. What can you do with it? You can extend the Linux kernel behavior for a variety of purposes such as load balancing, container networking, kernel tracing, monitoring, and security. You can solve those production issues where user-space solutions alone aren’t enough by executing the user-space code in the kernel. Btrfs Btrfs is a copy-on-write (CoW) filesystem, which means that instead of overwriting in one place, all the updates to metadata or file data are written to a new location on the disk. Btrfs mainly focuses on fault tolerance, repair, and easy administration. It supports features such as snapshots, online defragmentation, pooling, and integrated multiple device support. It is the only filesystem implementation that works with resource isolation. What can you do with it? You can address and manage large storage subsystems by leveraging features like snapshots, load balancing, online defragmentation, pooling, and integrated multiple device support. You can manage, detect, and repair errors with data and metadata checksums, mirroring, and file self-healing. Netconsd (Netconsole daemon) Netconsd is a UDP-based daemon that provides lightweight transport for Linux netconsole messages. It receives and processes log data from the Linux kernel and serves it up as a structured data. Simply put, it is a kernel module that sends all kernel log messages over the network to another computer, without involving user space. What can you do with it? Detect, reorder, or request retransmission of missing messages with the provided metadata. Extract meaningful signal from the data logged by netconsd to rapidly identify and diagnose misbehaving services. Cgroup2 Cgroup2 is a Linux kernel feature that allows you to group and structure workloads and also control the amount of system resources assigned to each group. It consists of controllers for memory, I/O, central processing unit, and more. Using cgroup2, you can isolate workloads, prioritize, and configure the distribution of resources. What can you do with it? You can create isolated groups of processes and then control and measure the distribution of memory, IO, CPU and other resources for each group. You can detect resource shortages using PSI pressure metrics for memory, IO, and CPU with cgroup2. With cgroup2, production engineers will be able to deal with increasing resource pressure more proactively and prevent conflicts between workloads. Along with these products, they have open-sourced Pressure Stall Information (PSI), oomd, and many others. You can find the complete list of these products at Facebook Open Source website and also check out the official announcement. Facebook open sources QNNPACK, a library for optimized mobile deep learning Facebook introduces two new AI-powered video calling devices “built with Privacy + Security in mind” Facebook’s Glow, a machine learning compiler, to be supported by Intel, Qualcomm and others

Facebook AI Research (FAIR) and NYU school of medicine announced yesterday that they're releasing new open source AI research models and data as a part of FastMRI. FastMRI is a new collaborative research project by Facebook and NYU School of medicine, that was announced back in August this year.   FastMRI makes use of artificial intelligence (AI) to make the (MRI) scans up to 10 times faster. By releasing these new AI models and the MRI data, the FastMRI team aims to help improve diagnostic imaging technology, which in turn can increase patients’ access to more powerful and life-saving technology. The latest release explores new AI models, and the first large-scale MRI data set for reconstructing MRI scans. Let’s have a look at these key releases. First large-scale database for MRI scans The fastMRI team has come out with baseline models for ML-based image reconstruction from k-space data subsampled at 4x and 8x scan accelerations. A common challenge faced by AI researchers in the field of MR reconstruction is consistency, as they use a variety of datasets for training AI systems. This is why the latest and the largest open source MRI dataset will help tackle this problem of MR image reconstruction by providing an industry-wide and benchmark ready dataset. This dataset comprises approximately 1.5 million MR images drawn from 10,000 scans, as well as raw measurement data from nearly 1,600 scans. NYU fully anonymized the data set, as well as the metadata and image content manually.  It includes the k-space data collected during scanning. NYU School of Medicine has decided to offer researchers with unprecedented access to data so that they can easily train their models, validate their performance, and get a general idea on how image reconstruction techniques could be used in real-world conditions. The k-space data in this data set is derived from MR devices comprising multiple magnetic coils. It also comprises data simulating the measurements from single-coil machines. AI models, baselines, and results leaderboard FastMRI team mainly focused on two tasks, namely, single-coil reconstruction and multi-coil reconstruction. In both the single-coil and multi-coil deep learning baselines, the AI models are based on u-nets, a convolutional network architecture developed specifically for image segmentation in biomedical applications. U-nets also has a proven track record with an image-to-image prediction. Moreover, a baseline for both classical and non-AI based reconstruction methods has been developed. A separate baseline comprising deep learning models has also been created. Apart from that, FAIR has created a leaderboard for the consistent measurement of MR progress and reconstruction results. The team has already added the baseline models to start with. Researchers can further add improved results as they begin generating and submitting the results to conferences and journals with the help of the fastMRI data set. It will also help the researchers in evaluating their results against the consistent metrics and to figure out how different approaches compare. “Our priority for the next phase of this collaboration is to use the experimental foundations we’ve established — the data and baselines — to further explore AI-based image reconstruction techniques. Additionally, any progress that we make at FAIR and NYU School of Medicine will be part of a larger collaboration that spans multiple research communities” says the FAIR team. For more information, check out the official blog post. Facebook AI researchers investigate how AI agents can develop their own conceptual shared language Facebook plans to change its algorithm to demote “borderline content” that promotes misinformation and hate speech on the platform Babysitters now must pass Perdictim’s AI assessment to be “perfect” to get the job

On 8th November, Mark Luckie, a former strategic partner manager for Facebook, posted an internal memo to Facebook Employees which opined how Facebook is “failing its black employees and its black users.” The memo was sent shortly before he left the company and just days after the New York Times report which threw Facebook under scrutiny for its leadership morals. Facebook and its ‘black people problem’ Mark Luckie, whose job was to handle the firm’s relationship with “influencers” focused on underrepresented voices, detailed a wide range of problems faced both, internally and externally, by the Black Community at Facebook. He pointed out that Blacks are some of the most engaged and active members of Facebook's 2.2 billion-member community- more specifically, 63 percent of African Americans use Facebook to communicate with their family, and 60 percent use it to talk to their friends once a day, compared to 53 and 54 percent of the total U.S. population respectively (according to Facebook’s own research). Yet, many are unable to find a "safe space" for dialogue on the platform, find their accounts suspended indefinitely and their content being removed without notice. Luckie’s memo states: “When determining where to allocate resources, ranking data such as followers, the greatest number of likes and shares, or yearly revenue are employed to scale features and products, the problem with this approach is Facebook teams are effectively giving more resources to the people who already have them. In doing so, Facebook is increasing the disparity of access between legacy individuals/brands and minority communities.” "Facebook can't engender the trust of its black users if it can't maintain the trust of its black employees." In the memo, Luckie congratulated the tech giant for increasing the number of black employees from 2 percent to 4 percent in 2018.  That being said, he went to list down the many issues faced by employees and criticized the firm's human resources department for protecting managers instead of supporting employees in lieu of such incidents. He said, "I've heard far too many stories from black employees of a colleague or manager calling them "hostile" or "aggressive" for simply sharing their thoughts in a manner not dissimilar from their non-black team members, a few black employees have reported being specifically dissuaded by their managers from becoming active in the [internal] Black@ group or doing "black stuff," even if it happens outside of work hours." He pointed out the hypocrisy in the firm where buildings are covered with ‘Black Lives Matter' posters compared to actually appointing more black employees. The existing black employees are often hassled by security and viewed with suspicion by fellow employees. “To feel like an oddity at your own place of employment because of the color of your skin while passing posters reminding you to be your authentic self feels in itself inauthentic” He claimed that Black staffers at Facebook subdue their voices for the fear of risking or jeopardizing their professional relationships and career advancement. After-effects of Mark’s memo Mr Luckie’s comments created waves around social media. What followed was a pattern we are all familiar with: ‘deny and deflect the blame’. First came the public statement, from Facebook spokesman Anthony Harrison: “Over the last few years, we’ve been working diligently to increase the range of perspectives among those who build our products and serve the people who use them throughout the world. The growth in the representation of people from more diverse groups, working in many different functions across the company, is a key driver of our ability to succeed, we want to fully support all employees when there are issues reported and when there may be micro-behaviors that add up. We are going to keep doing all we can to be a truly inclusive company.” As reported by BBC news, the statement was followed by an internal leak, that while Mr Luckie’s post was made public on Tuesday, it had been circulated at Facebook on 8th  November. At that time, Ime Archibong, Facebook’s director of product partnerships responded to the memo. On Tuesday, Mr Luckie posted his response on Twitter, suggesting Facebook’s tone publicly did not necessarily match what was said to him internally. https://twitter.com/marksluckie/status/1067494650259345408 Mr. Luckie seemed to attempt to protect Mr. Archibong’s identity, however, missed out an ‘Ime’ in his tweet. Mr. Archibong- who is also black- has confirmed he wrote the comments. https://twitter.com/_ImeArchibong/status/1067520926114148352 He was disappointed that the conversation was made public, and described Mr Luckie’s note as “pretty self-serving and disingenuous” while accusing him of having a “selfish agenda and not one that has the best intentions of the community and people you likely consider friends at heart”. The whole situation again suggests that Facebook is more concerned with not looking bad, rather than assessing if it is doing bad and what can it do to make its forum more approachable and safe for different members of the community. Mark’s Recommendations to “improve Facebook’s relationship with diverse communities” Mark ends the memo with some recommendations for the company, some of these include: For any team that has one or more people dedicated specifically to diversity, require a strategic plan for how that work will be incorporated into larger goals for the team. Create metrics for other team members to incorporate into their goals as well that ensure representation is everyone's responsibility. Implement data-driven goals to ensure partnerships, product testing, and client support is reflective of the demographics of Facebook. Level up cultural competency training for Operations teams that review reported infractions on Facebook and Instagram. Whenever possible, avoid relying solely on algorithms or AI to triage these problems. Create internal systems for employees to anonymously report microaggressions. This includes using coded language like “lowering the bar” or “hostile,” disproportionately giving lower performance review scores to women and people of color, or discouraging employees from engaging in cultural activities outside of their agreed upon work schedule. If these reported infractions surface a pattern, require the manager and/or team to attend sensitivity training to amend the behavior. Support emerging talent and brands by creating a pipeline of communication and scaled support that allows them to further build with the platform. Establish more regularly-scheduled focus groups with underrepresented communities, particularly the Black and Latino users who over-engage on Facebook and Instagram. Use these conversations to gain insight on how to grow the platform. After Marks memo went viral, many black employees from big tech companies came forward with their own stories of harassment at the workplace, including athlete Leslie Miller who tweeted: https://twitter.com/shaft/status/1067479669593726976 The memo's publication comes on the same day that a Facebook executive was grilled by parliamentary leaders from nine different countries at a special hearing on disinformation in the United Kingdom. You can head over Facebook’s Blog to read the memo in its entirety. NYT Facebook exposé fallout: Board defends Zuckerberg and Sandberg; Media call and transparency report Highlights Outage plagues Facebook, Instagram and Whatsapp ahead of Black Friday Sale, throwing users and businesses into panic Facebook’s outgoing Head of communications and policy takes blame for hiring PR firm ‘Definers’ and reveals more

On Tuesday, Adam Gaier, a student researcher and David Ha, Staff Research Scientist at the Google research team published a paper on Weight Agnostic Neural Networks (WANN) that can perform tasks even without learning the weight parameters. In “Weight Agnostic Neural Networks”, researchers present their first step towards searching networks with the neural net architectures that can already perform tasks, even when they use a random shared weight. The team writes, “Our motivation in this work is to question to what extent neural network architectures alone, without learning any weight parameters, can encode solutions for a given task. By exploring such neural network architectures, we present agents that can already perform well in their environment without the need to learn weight parameters.” The team looked at analogies of nature vs. nurture. They gave an example of certain precocial species in biology—who possess anti-predator behaviors from the moment of birth and can perform complex motor and sensory tasks without learning, Hence, researchers constructed network architectures that can perform well without training. The team has also open-sourced the code to reproduce WANN experiments in the broader research community. Researchers explored range of WANNs using topology search algorithm The team started with a population of minimal neural network architecture candidates, which have very few connections, and used a well-established topology search algorithm to evolve the architecture by adding single connections and single nodes one by one. Unlike traditional neural architecture search methods, where all of the weight parameters of new architectures need to be trained using a learning algorithm, the team took a simpler approach. To all candidate architectures the team first assigned a single shared weight value at each iteration, and then optimized to perform well over a wide range of shared weight values. In addition to exploring a range of weight agnostic neural networks, researchers also looked for network architectures that were only as complex as they need to be. They accomplished this by optimizing for both the performance of the networks and their complexity simultaneously, using techniques drawn from multi-objective optimization. Source: Google AI blog. Overview of Weight Agnostic Neural Network Search and corresponding operators for searching the space of network topologies. Training the WANN architectures Researchers believe that unlike traditional networks, WANNS can be easily trained by finding the best single shared weight parameter that maximizes its performance. They proved this with an example of a swing-up cartpole task using constant weights:   Source: Google AI blog. A WANN performing a Cartpole Swing-up task at various different weight parameters and fine tune weights As per the above figure, WANNs can perform tasks using a range of shared weight parameters. However, the performance is not comparable to a network that learns weights for each individual connection, normally done in network training. To improve performance, researchers used the WANN architecture, and the best shared weight to fine-tune the weights of each individual connection using a reinforcement learning algorithm, like how a normal neural network is trained. Created an ensemble of multiple distinct models of WANN architecture The researchers also believe that by using copies of the same WANN architecture, where each copy of the WANN is assigned a distinct weight value, they created an ensemble of multiple distinct models for the same task. And according to them this ensemble generally achieves better performance than a single model. They illustrated this with an example of an MNIST classifier: Source: Google AI blog The team conclude that a conventional network with random initialization will achieve ~10% accuracy on MNIST. While this particular network architecture that uses random weights when applied to MNIST achieves an accuracy of > 80%. However, when an ensemble of WANNs is used the accuracy increases to > 90%. The researchers hope that this work will serve as a stepping stone to discover novel fundamental neural network components such as convolutional networks in deep learning. To know more about this research, check out the official Google AI blog. What’s new in machine learning this week? DeepMind introduces OpenSpiel, a reinforcement learning-based framework for video games NVIDIA’s latest breakthroughs in conversational AI: Trains BERT in under an hour, launches Project Megatron to train transformer based models at scale Google open sources an on-device, real-time hand gesture recognition algorithm built with MediaPipe

This could be Amazon’s yet another shot across the bow to Oracle. The undisputed cloud king is well aware its database segment is a small fish in a pond dominated by Oracle. But as more number of enterprises move from on-premise to the cloud, Amazon's database market share could improve. One of the standout announcement from this year’s re:Invent conference was a “serverless database” based on and expanding upon the company’s fully managed Aurora database architecture. Aurora Serverless will let customers create database instances that only run when needed and automatically scale up or down based on demand. If a database isn’t needed at all, it will shut down until it is needed. This way, users will be able to pay by the second for the Aurora Serverless computation that they use – they won’t end up footing the bill for a database sitting idle overnight. In essence, Aurora was itself a pretty good database model in itself, in an environment where the workload was predictable. But the Amazon Web Services (AWS) eventually realized the workloads can be intermittent in some cases, and equally unpredictable at other times as requests may arrive in a span of few minutes or hours per day or per week. This is where the new variant of Aurora comes into the picture. Aurora Serverless has been designed keeping in mind workloads that are highly variable and subject to rapid change. Further, you are paying on a second-by-second basis, for the actual database resources you use. “Because storage and processing are separate, you can scale all the way down to zero and pay only for storage. I think this is really cool,” AWS evangelist Jeff Barr said, describing the serverless model that builds on a clean separation of processing and storage (an intrinsic part of the Aurora architecture). So in use cases when you have a low-volume blog site which is only used for a few minutes several times per day or week, or applications which peak for around 30 minutes each day or several times per year such as the HR budgeting and operational reporting forms, Aurora Serverless auto-scales to the capacity requirements. There could also be cases when the peak of activity is hard to predict, such as a traffic site which may get all of a sudden ‘active’ when it starts raining. Here again, the serverless database meets the needs of peak load, and then scales back down when the surge is over. This is a rather upright feature that has been introduced. Your developers may be using databases during work hours but they certainly don’t need them on nights or weekends. Thanks to Aurora Serverless, your database automatically shuts down when not in use. On the other hand, manually managing database capacity for each application is not a sensible approach – it can take up valuable time and lead to inefficient use of database resources. With Aurora Serverless, you simply create a database endpoint, optionally specify the desired database capacity range, and connect your applications. The endpoint is a simple proxy that routes your queries to a rapidly scaled fleet of database resources. This allows your connections to remain intact without disruptions, even as scaling operations take place behind the scenes. You can also migrate between standard and serverless configurations with a few clicks in the AWS Management Console. Such  an on-demand, auto-scaling configuration where the database automatically starts up, shuts down, and scales up or down capacity based on your application's needs, Aurora Serverless is truly how you ‘reinvent’ an Aurora database. 2018 will make it more clear how the new database is actually implemented. Meanwhile, you can sign up for the preview of Aurora Serverless by filling up this form.

Microsoft made a series of new announcements, earlier this week. These include a new Microsoft Defender ATP for Mac, a first fully automated DNA data storage system, and the Revived Microsoft Office Assistant, Clippy. Microsoft Defender ATP for Mac Microsoft team announced yesterday that it's expanding the reach of the core components of its security platforms (including the new Threat & Vulnerability Management) to Mac devices. Also, the name of these unified endpoint security platforms has been updated to Microsoft Defender ATP (Advanced Threat Protection) from the prior Windows Defender ATP, keeping in mind its new cross-platform nature. “We’ve been working closely with industry partners to enable Windows Defender Advanced Threat Protection (ATP) customers to protect their non-Windows devices while keeping a centralized “single pane of glass” experience”, states the Microsoft Team. Users can install the Microsoft Defender ATP client on devices running macOS Mojave, macOS High Sierra, or macOS Sierra to manage and protect these devices. This app offers next-gen anti-malware protection, allowing users to review and perform configuration of their protection. Users can also configure the advanced settings, including disabling or enabling real-time protection, cloud-delivered protection, and automatic sample submission among others. Moreover, devices with alerts and detections will also get surfaced in the Microsoft Defender ATP portal. Security analysts and admins can then further review these alerts on Mac devices. Other than that, the Microsoft team also plans to bring Microsoft Intune in the future. This would enable the users to configure and deploy the settings via alternative Mac and MDM management tools such as JAMF. Fully automated DNA data storage system Microsoft announced the new and first fully automated DNA data storage system, yesterday. The system allows with the storage and retrieval of data in manufactured DNA. This move is aimed at moving the DNA tech out of the research lab and into commercial data centers, says the Microsoft team. The team (Microsoft researchers and University of Washington) successfully encoded the word “hello” in snippets of fabricated DNA. They then further converted it back to digital data with the help of a fully automated end-to-end system. This automated DNA data storage system makes use of the software developed by the Microsoft and UW team that helps convert the ones and zeros of digital data into the As, Ts, Cs, and Gs (the building blocks of DNA). It then leverages the inexpensive, ‘off-the-shelf’  lab equipment to allow the flow of necessary liquids and chemicals into a synthesizer. This synthesizer then builds the manufactured snippets of DNA and pushes them into a storage vessel. In case the system wants to retrieve the information, it can add other chemicals to properly prepare the DNA and uses microfluidic pumps to push the liquids into other parts of the system. This system is then able to “read” the DNA sequences and convert them back to information understandable by a computer. According to the researchers, “the goal of the project was not to prove how fast or inexpensively the system could work, but simply to demonstrate that automation is possible” Revived Office Assistant Clippy Microsoft revived its 90s Microsoft Office Assistant, called Clippy, earlier this week on Tuesday. Microsoft Office team brought back Clippy as an app that can offer animated Clippy stickers on chats in Microsoft Teams, company’s group chat software.These Clippy stickers were also released on Microsoft’s official Office developer GitHub page, allowing all the Microsoft Teams users to import and use these stickers for free. However, Clippy was removed yet again the next day. This is because the “brand police” within Microsoft was not happy with the reappearance of Clippy on Microsoft Teams, reports The Verge. The GitHub project associated with the same has also been removed. Clippy fans, however, are not happy with the company’s decision and have started a thread requesting Microsoft to bring back Clippy in Microsoft Teams. Microsoft brings PostgreSQL extension and SQL Notebooks functionality to Azure Data Studio Microsoft open-sources Project Zipline, its data compression algorithm and hardware for the cloud Microsoft announces Game stack with Xbox Live integration to Android and iOS