Tech News - Data

Structure Query Language (SQL), which was once considered “not a serious language” by its authors, has now become a dominant query language for relational databases in the industry. Its battle-tested solutions, stability, portability makes it a reliable choice to perform operations on your stored data. However, it does has its share of weak points and that’s what Elvis Pranskevichus, founder of EdgeDB, listed down in a post titled “We Can Do Better Than SQL” published yesterday. He explained that we now need a “better SQL” and further introduced the EdgeQL language, which aims to address the limitations in SQL. SQL’s shortcomings Following are some of the shortcomings Pranskevichus talks about in his post: “Lack of Orthogonality” Orthogonality is a property, which means if you make some changes in one component, it will have no side effect on any other component. In the case of SQL, it means, allowing users to combine a small set of primitive constructs in a small number of ways. Orthogonality leads to a more compact and consistent design and not having it will lead to language which has many exceptions and caveats. Giving an example, Pranskevichus wrote, “A good example of orthogonality in a programming language is the ability to substitute an arbitrary part of an expression with a variable, or a function call, without any effect on the final result.” SQL does not permit such type of generic substitution. “Lack of Compactness” One of the side effects of not being orthogonal is lack of compactness. SQL is also considered to be “verbose” because of its goal of being an English-like language for catering to “non-professions”. “However, with the growth of the language, this verbosity has contributed negatively to the ability to write and comprehend SQL queries. We learnt this lesson with COBOL, and the world has long since moved on to newer, more succinct programming languages. In addition to keyword proliferation, the orthogonality issues discussed above make queries more verbose and harder to read,” wrote Pranskevichus in his post. “Lack of Consistency” Pranskevichus further adds that SQL is inconsistent in terms of both syntax and semantics. Additionally, there is a problem of standardization as well as different database vendors implement their own version of SQL, which often end up being incompatible with other SQL variants. Introducing EdgeQL With EdgeQL, Pranskevichus aims to provide users a language which is orthogonal, consistent, and compact, and at the same time works with the generally applicable relational model. In short, he aims to make SQL better! EdgeQL basically considers every value a set and every expression a function over sets. This design of the language allows you yo factor any part of an EdgeQL expression into a view or a function without changing other parts of the query. It has no null and a missing value is considered to be an empty set, which comes with the advantage of having only two boolean logic states. Read Pranskevichus’s original post for more details on EdqeQL. Building a scalable PostgreSQL solution PostgreSQL security: a quick look at authentication best practices [Tutorial] How to handle backup and recovery with PostgreSQL 11 [Tutorial]

Following the U.S President Donald Trump’s accusatory tweet on Tuesday morning, Google released a statement on the same day denying that it's algorithms favor liberal media outlets over right-wing ones. President Donald Trump went on to claim that Google search results for “Trump News” reports fake news. He accused the search engines’ algorithms of being rigged. Source: Twitter   Source: Twitter The 96% statistics were apparently taken from the results of a PJ Media investigation into Google News searches for the word "Trump".  The news came with a headline "96 Percent of Google Search Results for ‘Trump’ News Are from Liberal Media Outlets." Writer Paula Bolyard said she made the assessment after typing “Trump News” into Google’s ‘News’ tab across multiple computers, and then analyzed the top results against conservative journalist Sharyl Attkisson’s media bias chart. Trump has been tweeting since late July about discriminatory practices on Twitter and other social media sites more broadly and now his focus on Google is making rounds on the internet. Google’s spokesperson Riva Sciuto, addressed these accusations by stating that hundreds of improvements are done to the search giant’s algorithms each year to ensure that they surface high-quality content and the most relevant answers in response to users' queries. Google testifies that setting a political agenda was never entertained nor were its search results biased toward any political ideology These allegations- based on an analysis from the conservative online media outlet has little evidence behind them. “Google and Twitter and Facebook, they’re really treading on very, very troubled territory. And they have to be careful,” the president said later on Tuesday. “It’s not fair to large portions of the population.” Just 9 hours ago, the President posted another video with the caption #StopTheBias claiming that Google had promoted Barack Obama's state of the union address on its homepage but refused to do the same for Trump once he got elected as the president. The Trump administration on Tuesday said it might explore regulating Google, an effort that would challenge protections around free speech online. Jack Dorsey to testify explaining Twitter algorithms before the House Energy and Commerce Committee Facebook, Twitter takes down fake accounts with ties to Russia and Iran, suspected to influence the US midterm elections Epic games CEO calls Google “irresponsible” for disclosing the security flaw in Fortnite Android Installer before patch was ready    

After announcing Redis 5 RC1 in May earlier this year, the stable version of Redis 5 was released yesterday. This release comes with a new Stream data type, LFU/LRU info in RDB, active defragmentation version 2, HyperLogLogs improvements and many other improvements. What is new in Redis 5? Redis 5 comes with a new data type called Stream, which models a log data structure in a more abstract way. Three new modules got important APIs: Cluster API, Timer API, Dictionary API. With these APIs, you can now build a distributed system with Redis using it just as a framework, creating your own protocols. To provide better-caching accuracy after a restart or when a slave does a full sync, RDB now stores the LFU and LRU information. In the future releases, we are likely to see a new feature that sends TOUCH commands to slaves to update their information about hot keys. The cluster manager is now ported from Ruby to C and is integrated with redis-cli. Because of this change, it is faster and no longer has any dependency. To learn more about the cluster manager, you can run the redis-cli --cluster help command. Also, many commands with subcommands have a HELP subcommand. Sorted set commands, ZPOPMIN/MAX, and blocking variants are introduced. These commands are used in applications such as time series and leaderboards. With active defragmentation version 2, the process of defragmenting the memory of a running server is better than before. This will be very useful for long-running workloads that tend to fragment Jemalloc. Jemalloc is now upgraded to version 5.1 Improvements are made in the implementations of the HyperLogLog data structure with refined algorithms to offer a more accurate cardinality estimation. This version comes with better memory reporting capabilities. Redis 5 provides improved networking especially related to emitting large objects, CLIENT UNBLOCK and CLIENT ID for useful patterns around connection pools and blocking commands. Read the full Redis 5 release notes on GitHub. MongoDB switches to Server Side Public License (SSPL) to prevent cloud providers from exploiting its open source code Facebook open sources LogDevice, a distributed data store for logs RxDB 8.0.0, a reactive, offline-first, multiplatform database for JavaScript released!

Yesterday, the team at Facebook’s AI research announced that they have expanded and enhanced their LASER (Language-Agnostic SEntence Representations) toolkit to work with more than more than 90 languages, written in 28 different alphabets. This has accelerated the transfer of natural language processing (NLP) applications to many more languages. The team is now open-sourcing LASER and making it as the first exploration of multilingual sentence representations. Currently 93 languages have been incorporated into LASER. LASER achieves the results by embedding all languages together in a single shared space. They are also making the multilingual encoder and PyTorch code freely available and providing a multilingual test set for more than 100 languages. The Facebook post reads, “The 93 languages incorporated into LASER include languages with subject-verb-object (SVO) order (e.g., English), SOV order (e.g., Bengali and Turkic), VSO order (e.g., Tagalog and Berber), and even VOS order (e.g., Malagasy).” Features of LASER Enables zero-shot transfer of NLP models from one language, such as English, to scores of others including languages where training data is limited. Handles low-resource languages and dialects. Provides accuracy for 13 out of the 14 languages in the XNLI corpus. It delivers results in cross-lingual document classification (MLDoc corpus). LASER’s sentence embeddings are strong at parallel corpus mining which establishes a new state of the art in the BUCC, 2018 workshop on building and using comparable Corpora, shared task for three of its four language pairs. It provides fast performance with processing up to 2,000 sentences per second on GPU. PyTorch has been used to implement the sentence encoder with minimal external dependencies. LASER supports the use of multiple languages in one sentence. LASER’s performance improves as new languages get added and the system keeps on learning to recognize the characteristics of language families. Sentence embeddings LASER maps a sentence in any language to a point in a high-dimensional space such that the same sentence in any language will end up in the same neighborhood. This representation could also be a universal language in a semantic vector space. The Facebook post reads, “We have observed that the distance in that space correlates very well to the semantic closeness of the sentences.” The sentence embeddings are used for initializing the decoder LSTM through a linear transformation and are also concatenated to its input embeddings at every time step. The encoder/decoder approach The approach behind this project is based on neural machine translation, an encoder/decoder approach which is also known as sequence-to-sequence processing. LASER uses one shared encoder for all input languages and a shared decoder for generating the output language. LASER uses a 1,024-dimension fixed-size vector for representing the input sentence. The decoder is instructed about which language needs to be generated. As the encoder has no explicit signal for indicating the input language, this method encourages it to learn language-independent representations. The team at Facebook AI-research has trained their systems on 223 million sentences of public parallel data, aligned with either English or Spanish. By using a shared BPE vocabulary trained on the concatenation of all languages, it was possible to benefit  low-resource languages from high-resource languages of the same family. Zero-shot, cross-lingual natural language inference LASER achieves excellent results in cross-lingual natural language inference (NLI). The Facebook’s AI research team considers the zero-shot setting as they train the NLI classifier on English and then apply it to all target languages with no fine tuning or target-language resources. The distances between all sentence pairs are calculated and the closest ones are selected. For more precision, the margin between the closest sentence and the other nearest neighbors is considered. This search is performed using Facebook’s FAISS library. The team outperformed the state of the art on the shared BUCC task by a large margin. The team improved the F1 score from 85.5 to 96.2 for German/English, from 81.5 to 93.9 for French/English, from 81.3 to 93.3 for Russian/English, and from 77.5 to 92.3 for Chinese/English. To know more about LASER, check out the official post by Facebook. Trick or Treat – New Facebook Community Actions for users to create petitions and connect with public officials Russia opens civil cases against Facebook and Twitter over local data laws FTC officials plan to impose a fine of over $22.5 billion on Facebook for privacy violations, Washington Post reports

Last week, The US Government Accountability Office (GAO) released a report on developing an internet data privacy legislation to enhance consumer protections; the one similar to the EU's General Data Protection Regulation (GDPR). The GAO report was requested by the House Energy and Commerce Committee two years ago and have scheduled a hearing for February 26. During this hearing, the committee will discuss GAO’s findings and the possibility of drafting the US' first federal-level internet privacy law. GAO officials said, “Recent developments regarding Internet privacy suggest that this is an appropriate time for Congress to consider comprehensive Internet privacy legislation.” The GAO officials recommended that the Federal Trade Commission (FTC) should be put in charge of overseeing internet privacy enforcement. GAO investigators cited the Facebook Cambridge Analytica scandal as an example of why a federal-level internet privacy law is important. According to ZDNet, some of the examples include: The dangers to user privacy due to the lack of regulation and oversight in the ever-growing Internet of Things (IoT) sector where devices collect massive amounts of information without users' knowledge. Automakers collecting data from smart cars owners. The lack of federal oversight over companies that collect and resell user information. The lack of protections for mobile users against secret data collection practices. House Energy and Commerce Chairman Frank Pallone, Jr. (D-NJ), the official who requested the report in 2017, said, “This detailed GAO report makes clear that now is the time for comprehensive congressional action on privacy that should include ensuring any agency that oversees consumer privacy has the tools to protect consumers. These recommendations and findings will be helpful as we look to develop privacy legislation in the coming months.” For its report, the GAO committee also analyzed the FTC's previous 101 user internet privacy investigations. It also took into consideration feedback from the private sector, academia, advocacy groups, other government agencies, and nine former FTC and FCC top-ranking officials, including seven former commissioners. To know more about this news in detail, read the complete GAO report. U.S Government Accountability Office (GAO) reports U.S weapons can be easily hacked GDPR complaint claims Google and IAB leaked ‘highly intimate data’ of web users for behavioral advertising French data regulator, CNIL imposes a fine of 50M euros against Google for failing to comply with GDPR

At a press conference yesterday, French Economy Minister Bruno Le Maire announced that France would levy a new tax on big tech companies including Google, Apple, Facebook, and Amazon, also known as GAFA, w.e.f. January 1, 2019. This tax is estimated to bring in €500m, or about $567 million in the coming year. Le Maire told France24 television, “I am giving myself until March to reach a deal on a European tax on the digital giants. If the European states do not take their responsibilities on taxing the GAFA, we will do it at a national level in 2019.” In an interview with Reuters and a small group of European newspapers, Le Maire said, “We want a fair taxation of digital giants that creates value in Europe in 2019”. France, along with Germany’s help had proposed a comprehensive digital services tax (DST) to cover all 28 EU member states. However, Ireland dismissed the move stating that this would aggravate the US-EU trade intentions. Dublin also said this bloc should happen only after the Organisation for Economic Co-operation and Development (OECD) had presented its tax proposals in 2019. Le Maire, however, said that France would press ahead alone with the tax. In March 2018, the European Commission published a proposal for a 3% tax on tech giants with global revenues north of €750m ($850 million USD) per year, and EU revenue above €50m (about $57 million). But with disagreements by some member states, including Ireland and the Netherlands, on how to move forward with such a tax, the process has been stalled. Per Le Maire, “The digital giants are the ones who have the money." The companies "make considerable profits thanks to French consumers, thanks to the French market, and they pay 14 percentage points of tax less than other businesses.” In October, British Chancellor Philip Hammond announced in the Budget that he plans to introduce a digital services tax from April 2020 following a consultation. The Chancellor's office has suggested that the tax would generate at least 400 million pounds ($505 million) per year. According to Reuters, “President Emmanuel Macron’s government has proposed taxing the tech giants on revenues rather than profits, to get around the problem that the companies shift the profits from where they are earned to low tax jurisdictions.” France and Germany in their alternative plan at a meeting of EU finance ministers proposed levying a 3 percent tax on digital advertising from Google and Facebook, which together account for about 75 percent of digital advertising, starting in 2021. Ministers asked the European Commission to work on the new proposal and present its findings to them in January or February. After the meeting, Le Maire said, "It's a first step in the right direction, which in the coming months should make the taxation of digital giants a possibility." To know more about this in detail, visit France24’s complete coverage. Australia’s Assistance and Access (A&A) bill, popularly known as the anti-encryption law, opposed by many including the tech community Amazon addresses employees dissent regarding the company’s law enforcement policies at an all-staff meeting, in a first Senator Ron Wyden’s data privacy law draft can punish tech companies that misuse user data  

Deloitte launched their tenth annual “Tech Trends 2019: Beyond the digital frontier” report earlier this month. The report covers predictions related to Artificial intelligence, digital future, cloud, networking, and cybersecurity. Let’s have a look at these key predictions made in the report. More companies to transform into AI-fueled organizations Deloitte 2019 report states that there’ll be an increase in the number of companies following the transformation to fully autonomous AI-fueled firms in the next 18 to 24 months, making AI a major part of their corporate strategy. AI, machine learning, and other cognitive technologies run at the center of business and IT operations in an AI-fueled firm to harness data driven-insights. As per the two consecutive Deloitte global surveys (2016–17 and 2018), cognitive technologies/AI were at the top in a list of emerging technologies in which CIOs plan to invest. AI ambitions of these CIOs is more about using AI to increase productivity, and strengthen regulatory compliance using automation. Companies to make the transition from traditional to serverless environments (NoOps) The report states many CIOs will be looking at creating a NoOps IT environment that is automated and abstracted from the underlying infrastructure. It requires small teams to manage it and will thereby allow the CIOs to invest larger human capacity in developing new capabilities that can improve the overall operational efficiency. In NoOps environments, traditional operations like the code deployment and patching schedules are internal responsibilities and are mainly automated. This shift from traditional to serverless computing will allow the cloud vendors to dynamically and automatically allocate the compute, storage, and memory depending on the request for a higher-order service. Traditional cloud service models required organizations to manually design and provision such allocations. Serverless computing offers limitless scalability, high availability, NoOps, along with zero idle time costs. More companies are expected to take advantage of advanced connectivity to configure and operate enterprise networks As per the Deloitte report, many companies will opt for advanced networking to drive the development of new products and to transform inefficient operating models. CIOs are going to be virtualizing parts of the connectivity stack with the help of network management techniques like Software-defined networking (SDN) and network function virtualization (NFV). SDN is primarily used in data centers but its use is now being extended for wide area networking to connect data centers. NFV replaces network functions such as routing, switching, encryption, firewalling, WAN acceleration, etc, and can scale horizontally or vertically on demand. The report states that enterprises will be able to better optimize or “spin up” the network capabilities on demand to fulfill the needs of a specific application or meet the end-user requirements. Growth in interfaces like computer vision, gesture control devices, etc, will transform how humans, machines, and data interact with each other The report states that although conversational technologies are currently dominating the intelligent interfaces arena, other new additions in interfaces such as computer vision, gesture control devices, embedded eye-tracking platforms, bioacoustic sensing, and emotion detection/recognition technology, are gaining ground. Intelligent interfaces help track customers’ offline habits, similar to how search engines and social media companies track their customers’ digital habits. These interfaces also help understand the customers at a personal, and more detailed level, making it possible to “micro-personalize” the products and services. We will see more of these new interfaces combined with leading-edge technologies in the future (such as machine learning, robotics, IoT, contextual awareness, advanced augmented reality, and virtual reality) to transform the way we engage with machines, data, and each other. CMOs and CIOs will partner up to elevate the human experience by moving beyond marketing The report states that channel-focused services such as websites, social and mobile platforms, content management tools, search engine optimization are slowly becoming a thing of the past.  Many organizations will be moving beyond marketing by adopting a new generation of Martech systems and approach to data gathering, decision-making (determining how and when to provide an experience), and delivery (consistent delivery of dynamic content across channels). This, in turn, helps companies create personalized and dynamic end-to-end experiences for users and builds deep emotional connections among users to products and brands. Also, CMOs are required to own the delivery of the entire customer experience, and they often find themselves almost playing the CIO’s traditional role. At the same time, CIOs are required to transform the legacy systems and come out with new infrastructure to support the next-generation data management and customer engagement systems. This is why CIOs and CMOs will collaborate more closely to deliver on their company’s new marketing strategies as well as on the established digital agendas. Organizations to embed DevSecOps to improve cybersecurity As per the report, many organizations have started to use a method called DevSecOps that includes embedding security culture, practices, and tools into different phases of their DevOps pipelines. DevSecOps help improves the security and maturity levels of a company’s DevOps pipeline. DevSecOps is not a security trend but it's a new approach that offers companies a different way of thinking about security. DevSecOps has multiple benefits. It helps the security architects, developers, and operators share their metrics aligning to security and put a focus on business priorities. Organizations embedding DevSecOps into their development pipelines can use operational insights and threat intelligence. It also helps with proactive monitoring that involves automated, and continuous testing to identify problems. The report recommends that DevSecOps should tie in your broader IT strategy, which should be further driven by your business strategy. “If you can be deliberate about sensing and evaluating emerging technologies..you can make the unknown knowable..creating the confidence and construct to embrace digital while setting the stage to move beyond the digital frontier”, reads the report. For more information, check out the official Deloitte’s 2019 tech trends. IEEE Computer Society predicts top ten tech trends for 2019: assisted transportation, chatbots, and deep learning accelerators among others We discuss the key trends for web and app developers in 2019 [Podcast] We discuss the key trends for web and app developers in 2019 [Podcast]

Documents released this week, reveals that the data mining firm, Palantir was responsible for 2017 operation that targeted and arrested family members of children crossing the border alone. The documents show a huge contrast to what Palantir said its software was doing. This discrepancy was first identified by Mijente, an advocacy organization that has closely tracked Palantir’s murky role in immigration enforcement. The documents confirm that “the role Palantir technology played in facilitating hundreds of arrests, only a small fraction of which led to criminal prosecutions”, The Intercept reports. Palantir, a software firm founded by Peter Thiel, one of President Trump’s most vocal supporters in Silicon Valley, develops software that helps agents analyze massive amounts of personal data and builds profiles for prosecution and arrest. Also, in May 2018, Amazon employees, in a letter to Jeff Bezos, protested against the sale of its facial recognition tech to Palantir where they “refuse to contribute to tools that violate human rights”, citing the mistreatment of refugees and immigrants by ICE. Read Also: Amazon addresses employees dissent regarding the company’s law enforcement policies at an all-staff meeting, in a first Palantir earlier said it was not involved with the part of ICE, which was strictly devoted to deportations and the enforcement of immigration laws. Whereas Palantir’s $38 million contract with Homeland Security Investigations, or HSI, a component of ICE had a far broader criminal enforcement mandate. https://twitter.com/ConMijente/status/1124056308943138834 The 2017 ICE operation was designed to dissuade children from joining family members in the United States by targeting parents and sponsors for arrest. According to The Intercept, “Documents obtained through Freedom of Information Act litigation and provided to The Intercept show that this claim, that Palantir software is strictly involved in criminal investigations as opposed to deportations, is false.” As part of the operation, ICE arrested 443 people solely for being undocumented. For all this, Palantir’s software was used throughout, which helped agents build profiles of immigrant children and their family members for the prosecution and arrest of any undocumented person they encountered in their investigation. https://twitter.com/ConMijente/status/1124056314106322944 “The operation was underway as the Trump administration detained hundreds of children shelters throughout the country. Unaccompanied children were taken by border agents, sent to privately-run facilities, and held indefinitely. Any undocumented parent or family member who came forward to claim children were arrested by ICE for deportation. More children were kept in detention longer, as relatives stopped coming forward”, Mijente reports. Mijente further mentions in their post, “Mijente is urging Palantir to drop its contract with ICE and stop providing software to agencies that aid in tracking, detaining, and deporting migrants, refugees, and asylum seekers. As Palantir plans its initial public offering, Mijente is also calling on investors not to invest in a company that played a key role in family separation.” The seven-page document, titled “Unaccompanied Alien Children Human Smuggling Disruption Initiative,” details how one of Palantir’s software solutions, Investigative Case Management (ICM) can be used by agents stationed at the border to build cases of unaccompanied children and their families.Mijente further mentions, “This document is further proof that Palantir’s software directly aids in prosecutions for deportation carried out by HSI agents. Not only are HSI agents involved in deportations in the interior, but they are also actively aiding border agents by investigating and prosecuting relatives of unaccompanied children hoping to join their families.” Jesse Franzblau, senior policy analyst for the National Immigrant Justice Center, said in an email to The Intercept, “The detention and deportation machine is not only driven by hate, but also by profit. Palantir profits from its contract with ICE to help the administration target parents and sponsors of children, and also pays Amazon to use its servers in the process. The role of private tech behind immigration enforcement deserves more attention, particularly with the growing influence of Silicon Valley in government policymaking. “Yet, Palantir’s executives have made no move to cancel their work with ICE. Its founder, Alex Karp, said he’s “proud” to work with the United States government. Last year, he reportedly ignored employees who “begged” him to end the firm’s contract with ICE”, the Mijente report mentions. To know more about this news in detail head over to the official report. Lerna relicenses to ban major tech giants like Amazon, Microsoft, Palantir from using its software as a protest against ICE Free Software Foundation updates their licensing materials, adds Commons Clause and Fraunhofer FDK AAC license “We can sell dangerous surveillance systems to police or we can stand up for what’s right. We can’t do both,” says a protesting Amazon employee

There seems to be a running battle between Google and its employees for quite some time now. A group of Google employees announced yesterday that they’re launching a public awareness social media campaign from 9 AM to 6 PM EST today. The group, called, ‘Googlers for ending forced arbitration’ aims to educate people about the forced arbitration policy via Instagram and Twitter where they will also share their experiences about the same with the world.   https://twitter.com/endforcedarb/status/1084813222505410560 The group has researched their fellow tech employees, academic institutions, labour attorneys, advocacy groups, etc as well as the contracts of around 30 major tech companies, as a part of its efforts. They also published a post on Medium, yesterday, stating that “ending forced arbitration is the gateway change needed to transparently address inequity in the workplace”. According to National Association of Consumer Advocates, “In forced arbitration, a company requires a consumer or employee to submit any dispute that may arise to binding arbitration as a condition of employment or buying a product or service. The employee or consumer is required to waive their right to sue, to participate in a class action lawsuit, or to appeal”. https://twitter.com/ODwyerLaw/status/1084893776429178881 Demands for more transparency around Google’s sexual assault policies seems to have become a bone of contention for Google. For instance, two shareholders, namely, James Martin and two other pension funds sued Alphabet’s board members, last week, for protecting the top execs accused of sexual harassment. The lawsuit, which seeks major changes to Google’s corporate governance, also urges for more clarity surrounding Google’s policies. Similarly, Liz Fong Jones, developer advocate at Google Cloud platform, revealed earlier this month, that she’s planning to leave the firm due to Google showing lack of leadership in case of the demands made by employees during the Google walkout. It was back in November 2018 when over 20,000 Google employees organized Google “walkout for real change” and walked out of their offices along with temps and contractors to protest against the discrimination, racism, and sexual harassment encountered within Google. Google employees had made five demands as part of the walkout, including ending forced arbitration for all employees (including temps) in cases of sexual harassment and other forms of discrimination. Now, although Google announced that it's ending its forced Arbitration policy as a response to the walkout (a move that was soon followed by Facebook) back in November, Google employees are not convinced. They argue that the announcement only made up for strong headlines, and did not actually do enough for the employees. The employees mentioned that there were “no meaningful gains for worker equity … nor any actual change in employee contracts or future offer letters (as of this publication, we have confirmed Google is still sending out offer letters with the old arbitration policy)”. Moreover, forced arbitration still exists in Google for cases involving other forms of workplace harassment and discrimination issues that are non-sexual in nature. Google has made the forced arbitration policy optional only for individual cases of sexual assault for full-time employees and still exists for class-action lawsuits and thousands of contractors who work for the company. Additionally, the employee contracts in the US still have the arbitration waiver in effect. “Our leadership team responded to our five original demands with a handful of partial policy changes. The other ‘changes’ they announced simply re-stated our current, ineffective practices or introduced extraneous measures that are irrelevant to bringing equity to the workplace”, mentions the group in a blog post on Medium. Follow the public awareness campaign on the group’s Instagram and Twitter accounts. Recode Decode #GoogleWalkout interview shows why data and evidence don’t always lead to right decisions in even the world’s most data-driven company Tech Workers Coalition volunteers talk unionization and solidarity in Silicon Valley BuzzFeed Report: Google’s sexual misconduct policy “does not apply retroactively to claims already compelled to arbitration”

The Google Brain team with DeepMind and ETH Zurich have introduced an episodic memory-based curiosity model which allows Reinforcement Learning (RL) agents to explore environments in an intelligent way. This model was the result of a study called Episodic Curiosity through Reachability, the findings of which Google AI shared yesterday. Why this episodic curiosity model is introduced? In real-world scenarios, the rewards required in reinforcement learning are sparse and most of the current reinforcement learning algorithms struggle with such sparsity. Wouldn’t it be better if the agent is capable of creating its own rewards? That’s what this model does. This makes the rewards denser and more suitable for learning. Many researchers have worked on some curiosity-driven learning approaches before, one of them is Intrinsic Curiosity Module (ICM). This method is explored in the recent paper Curiosity-driven Exploration by Self-supervised Prediction published by Ph.D. students at the University of California, Berkeley. ICM builds a predictive model of the dynamics of the world. The agent is rewarded when the model fails to make good predictions. Exploring unvisited locations is not directly a part of the ICM curiosity formulation. In the ICM method, visiting them is only a way to obtain more “surprise” and thus maximize overall rewards. As a result, in some environments there could be other ways to cause self-surprise, leading to unforeseen results. The authors of the ICM method along with researchers at OpenAI, in their research Large-Scale Study of Curiosity-Driven Learning, show a hidden danger of surprise maximization. Instead of doing something useful for the task at hand, agents can learn to indulge procrastination-like behavior. The episodic memory-based curiosity model overcomes this “procrastination” issues. What is episodic memory-based curiosity model? This model uses a deep neural network trained to measure how similar two experiences are. For training the model, the researchers made it guess whether two observations were experienced close together in time, or far apart in time. Temporal proximity is a good proxy for whether two experiences should be judged to be part of the same experience. This training gives a general concept of novelty via reachability, which is shown a follows: Source: Google AI How this model works Inspired by curious behavior in animals, this model rewards the agent with a bonus when it observes something novel. This bonus is summed up with the real task reward making it possible for the RL algorithm to learn from the combined reward. To calculate the bonus of the agent, the current observation is compared with the observation in memory. This comparison is done based on how many environment steps it takes to reach the current observation from those in memory. Source: Google AI This method follows these steps: The agent's observations of the environment are stored in an episodic memory. The agents are also rewarded for reaching observations that are not yet represented in memory. In this method, being “not in memory” is the definition of novelty in our method. Such a behavior of seeking the unfamiliar will lead the artificial agent to new locations, thus keeping it from wandering in circles and ultimately help it stumble on the goal. Experiment and results Different approaches to curiosity were tested in two visually rich 3D environments: ViZDoom and DMLab. The agent was given various tasks such as searching for a goal in a maze or collecting good objects and avoiding bad objects. The standard setting in previous formulations, such as ICM, on DMLab, was to provide the agent a laser-like science fiction gadget. If the agent does not need a gadget for a particular task, it was free not to use it. In this test, the surprise-based ICM method used this gadget a lot even when it is useless for the task at hand. The newly introduced method instead learns reasonable exploration behavior under the same conditions. This is because it does not try to predict the result of its actions, but rather seeks observations which are “harder” to achieve from those already in the episodic memory. In short, the agent implicitly pursues goals which require more effort to reach from memory than just a single tagging action. This approach penalizes an agent running in circles because after completing the first circle the agent does not encounter new observations other than those in memory, and thus receives no rewards. In the experimental environment, the model was able to achieve: In ViZDoom, the agent learned to successfully navigate to a distant goal at least two times faster than the state-of-the-art curiosity method ICM. In DMLab, the agent generalized well to new procedurally generated levels of the game. It was able to reach the goal at least two times more frequently than ICM on test mazes with very sparse reward. To know more in detail about the episodic memory-based curiosity model, check out Google AI’s post and also the paper: Episodic Curiosity Through Reachability. DeepMind open sources TRFL, a new library of reinforcement learning building blocks Google open sources Active Question Answering (ActiveQA), a Reinforcement Learning based Q&A system Understanding Deep Reinforcement Learning by understanding the Markov Decision Process [Tutorial]

Facebook employees are experiencing turbulent times as Apple decides to revoke the social media giant’s developer certificates. This is due to a TechCrunch report that said Facebook paid 20$/month to users including teens to install the ‘Facebook research app” on their devices which allowed them to track their mobile and web browsing activities. Following the revoke, Facebook employees will not be able to access early versions of Facebook apps such as Instagram and Messenger, and many other activities such as food ordering, locating an area on the map and much more. Yesterday, Apple announced that they have shut down the Facebook research app for iOS. According to Apple, “We designed our Enterprise Developer Program solely for the internal distribution of apps within an organization. Facebook has been using their membership to distribute a data-collecting app to consumers, which is a clear breach of their agreement with Apple”. The company further said, “Any developer using their enterprise certificates to distribute apps to consumers will have their certificates revoked, which is what we did in this case to protect our users and their data.” Per Mashable report, “Facebook employees argued that Apple's move was merely an attempt to distract from an embarrassing FaceTime bug that went public earlier in the week.” An employee commented, “Anything to take the heat off the FaceTime security breach.” Facebook also said that  it’s “working closely with Apple to reinstate our most critical internal apps immediately.” Mark Zuckerberg has also received a stern letter from Senator Mark Warner including a list of questions about the company’s data gathering practices, post the TechCrunch report went viral. In a statement, he mentioned, “It is inherently manipulative to offer teens money in exchange for their personal information when younger users don’t have a clear understanding of how much data they’re handing over and how sensitive it is.” Google disabled its iOS app too Similar to Facebook, Google too distributed a private app, Screenwise Meter, to monitor how people use their iPhones and rewarded the users with Google’s Opinion Rewards program gift cards in exchange for collecting information on their internet usage. However, yesterday, Google announced that it has disabled the iOS app. Google’s Screenwise Meter app has been a part of a program that’s been around since 2012. It first started tracking household web access through a Chrome extension and a special Google-provided tracking router. The app is open to anyone above 18 but allows users aged 13 and above to join the program if they’re in the same household. Facebook’s tracking app, on the other hand, targeted people between the ages of 13 and 25. A Google spokesperson told The Verge, “The Screenwise Meter iOS app should not have operated under Apple’s developer enterprise program — this was a mistake, and we apologize. We have disabled this app on iOS devices. This app is completely voluntary and always has been. We’ve been upfront with users about the way we use their data in this app, we have no access to encrypted data in apps and on devices, and users can opt out of the program at any time.” To know more about this news, head over to The Verge. Facebook researchers show random methods without any training can outperform modern sentence embeddings models for sentence classification Stanford experiment results on how deactivating Facebook affects social welfare measures Facebook pays users $20/month to install a ‘Facebook Research’ VPN that spies on their phone and web activities, TechCrunch reports

Adobe had already been venturing into the AI domain by coming up with products such as Adobe Sensei. Now, Adobe has developed a product that is said to be using Artificial Intelligence for detecting images that are heavily edited or images that have been tinkered with. Adobe is aiming to create more products in the AI space in order to build trust among people in digital media. Adobe has been widely used for editing images that express artistic creativity. However, some people use it to their own unfair advantage by manipulating images for deception. But, with AI in the game, the image deception problem seems to be getting fixed. A senior research scientist at Adobe, Vlad Morariu, has been working on computer vision technologies for a while now to detect manipulated images. Vlad mentions that there are existing tools which help trace digitally altered photos. For instance, different file formats have metadata that store information about the image captured and manipulated. Also, forensic tools help detect the altered images by analyzing strong edges, lighting, noise distribution, and pixel values of a photo. But, these tools are not as efficient at detecting fake images. Source: Adobe Vlad’s continuous research led him to come up with three new techniques in Artificial Intelligence for detecting image manipulation. Splicing: This combines different parts of two different images. Copy-move: This involves cloning or moving of objects within a photograph from one place to another. Removal: In this, an object within a photograph is removed and that space is filled in. This has greatly cut down on the time it would take forensic experts to detect fraud images. Vlad also mentions how they have trained a deep learning neural network to detect deception on thousands of known, manipulated images. It combines two different methods in one network to enhance the detection process even more. The first method makes use of RGB stream to detect tampering. And the second method uses a noise stream filter. Although these techniques are not foolproof, they provide more options for controlling digital manipulation currently. Adobe might get its hands dirty in the AI world even more in the future by including tools for detection of other kinds of manipulation in photographs. To know more about Adobe’s effort in controlling digital manipulation, check out Adobe’s official blog post. Adobe glides into Augmented Reality with Adobe Aero Adobe is going to acquire Magento for $1.68 Billion  

Researchers at MIT have come up with an intriguing approach to combat ‘Glioblastoma’- a malignant tumor of the brain/spinal cord- using machine learning techniques. By reducing the toxic chemotherapy and radiotherapy that is involved in treating this cancer, the researchers aim to improve the quality of life for patients, while also reducing the various side effects caused by the former using Reinforcement learning techniques. While the prognosis for adults is no more than 5 years, medical professionals try to shrink the tumor by administering drug doses in safe amounts. However, the pharmaceuticals are so strong that patients end up suffering from their side effects. Enter Machine Learning and Artificial Intelligence to save the day. While it's no hidden truth that machine learning is being incorporated into healthcare on a huge scale, the MIT researchers have taken this to the next level. Using Reinforcement Learning as the Big Idea to train the model Media Lab researcher Gregory Yauney will be presenting a paper next week at the 2018 Machine Learning for Healthcare conference at Stanford University. This paper details how the MIT Media Lab researchers have come up with a model that could make dosing cycles less toxic but still effective. Incorporating a “self-learning” machine-learning technique, the model studies treatment regimens being used presently, and iteratively changes the measurements. In the end, it finds an ideal treatment design suited to the patient. This has proven to reduce the tumor sizes to a degree almost identical to that of original medical regimens. The model simulated trials of 50 patients and designed treatments that either reduced dosages to twice a year or skipped them all together. This was done keeping in mind that the model has to shrink the size of the tumor but at the same time ensuring that reduced dosages did not lead to harmful side effects. The model is designed to used reinforced learning (RL)- that comprises artificially intelligent “agents” that complete “actions” in an unpredictable, complex environment to reach the desired outcome. The model’s agent goes through traditionally administered regimens. It uses a combination of the drugs temozolomide (TMZ) and procarbazine, lomustine, and vincristine (PVC), administered to the patients  over weeks or months. These regimens are based on protocols that have been used clinically for ages and are based on both, animal testing and various clinical tests and scenarios. The protocols are then used by Oncologists to predict how many doses the patients have to be administered based on weight. As the model explores the regimen, it decides on one of the two actions- Initiate a dose Withhold a dose If it does administer a dose, it has to make the decision if the patient needs the entire dose, or only a portion. After a decision is taken, the model checks with another clinical model to see if the tumor’s size has changed or if it’s still the same. If the tumor’s size has reduced, the model receives a reward else it is penalised. Rewards and penalties essentially are positive and negative numbers, say +1 or – 1. The researchers also had to ensure that the model does not over-dose or give out the maximum number of doses to reduce the mean diameter of the tumor. Therefore, the model is programmed in such a way that whenever it chooses to administer all full doses, it gets penalized. Thus the model is forced to administer fewer, smaller doses. Patik Shah, a principal investigator at the Media Lab who supervised this research, further stresses on the fact that, as compared to traditional RL models that work toward a single outcome, such as winning a game, and take any and all actions that maximize that outcome, the model implemented by the MIT researchers is a  “unorthodox RL model that weighs potential negative consequences of actions (doses) against an outcome (tumor reduction)” The model is strikingly wired to find a dose that does not necessarily maximize tumor reduction, but also establishes a perfect balance between maximum tumor reduction and low toxicity for the patients. The training and testing methodology used The model was trained on 50 simulated patients -  randomly selected from a large database of glioblastoma patients. These patients had previously undergone traditional treatments. The model conducted about 20,000 trial-and-error test runs for every patient. Once training was complete, the model understood the parameters for optimal regimens. The model was then tested on 50 new simulated patients and used the above-learned parameters to formulate new regimens based on various constraints that the researchers provided. The models treatment regimen was compared to the results of a conventional regimen using both TMZ and PVC. The outcome obtained was practically similar to the results obtained after the human counterparts administered treatments. The model was also able to treat each patient individually, as well as in a single cohort, and achieved similar results (medical data for each patient was available to the researchers). In short, the model has helped to generate precision medicine-based treatments by conducting one-person trials using unorthodox machine-learning architectures. Nicholas J. Schork, a professor and director of human biology at the J. Craig Venter Institute, and an expert in clinical trial design explains  “Humans don’t have the in-depth perception that a machine looking at tons of data has, so the human process is slow, tedious, and inexact,” he further adds  “Here, you’re just letting a computer look for patterns in the data, which would take forever for a human to sift through, and use those patterns to find optimal doses.” To sum it all up,  Machine learning is again proving to be an essential asset in the medical field- helping both researchers as well as patients to view medical treatments in an all new perspective. If you would like to know more about the progress done so far, head over to MIIT news. 23andMe shares 5mn client genetic data with GSK for drug target discovery Machine learning for genomics is bridging the gap between research and clinical trials 6 use cases of Machine Learning in Healthcare

Blockchain is an open decentralized database. It's the underlying technology for the popular cryptocurrency, Bitcoin. Now, banks and other financial institutions want to apply it to financial transactions. The recently launched IBM Blockchain platform, LedgerConnect is aimed at the financial industry and banking sectors. What is IBM blockchain? IBM, with CLS, a foreign exchange financial group, launched LedgerConnect a week ago. It is a proof of concept blockchain platform designed for companies that provide financial services. Its aim is to apply blockchain technology to a number of challenge areas that are currently not that fast. They also have challenges like tracking a paperwork trail, know your customer (KYC) processes, collateral management, and trades. So far, nine financial companies, including banks like Barclays and Citi, are involved in the trials. With everything being online, delays will likely be reduced by a large margin. According to an IBM report: “With IBM Blockchain, banks can create secure, low-cost and high volume cross-border payments without sacrificing margins.” Why is it important? The IBM website states that 91% of banks are investing in blockchain solutions by 2018. Also, 66% of institutions expect to be in production and running at scale with blockchain. All the transactions being on the same network makes everything much easier. An added advantage is that blockchain is known to be tamper-proof. So the overall security and speed will increase a lot. While the blockchain used in Bitcoin is public, the one used in large companies would be private. Makes sense as you wouldn’t want everyone’s banking and transaction info publically available everywhere. To further understand why blockchains are great for banking applications, read our article, 15 ways to make Blockchains scalable, secure and safe!. Impact of blockchain in banking and finance In the current day, banks work with manual transactions and a considerable amount of paperwork. So transactions like trades and loans take a lot of time, amounting to delays. Cross-border payments are costly and time-consuming. Moreover, making customers provide identification frequently is tedious and decreases customer satisfaction. IBM blockchain aims to address all these pain areas. As the platform is online, blockchain based smart-contracts automatically store information about the transactions in real-time. As a result, there won’t be discrepancies in the data shown in different locations. Transactions on blockchain are identified uniquely with a hash. KYC info will also be more secure since it will be stored in the blockchain that is known for being tamper-proof. Source: IBM infographic Source: IBM infographic LedgerConnect tries to create a network of multiple providers that can provide their applications to be deployed on this platform. Large banks and organizations can come and use those applications creating an app store like ecosystem. The applications in the LedgerConnect “store” are built on Hyperledger Fabric. This project is expected to launch in early 2019 but needs approvals from central banks. To know more about Hyperledger, take a look at our article Hyperledger: The Enterprise-ready Blockchain Implementing blockchain in finance and banking will enable simplicity and operational efficiency while also enhancing the end customer experience. Many of the current challenges can be solved and the time involved can be brought down to minutes from hours. To know more visit the IBM website. Read next: Google Cloud Launches Blockchain Toolkit to help developers build apps easily Oracle makes its Blockchain cloud service generally available Blockchain can solve tech’s trust issues – Imran Bashir

Earlier this year, Sundar Pichai shared seven AI principles that Google aims to follow in its work. Google also shared some best practices for building responsible AI. Yesterday, they shared the additional initiative and processes they have introduced to live up to their AI principles. Some of these initiatives include educating people about ethics in technology, introducing a formal review structure for new projects, products, and deals. Educating Googlers on ethical AI Making Googlers aware of the ethical issues: Additional learning material has been added to the Ethics in Technology Practice course that teaches technical and non-technical Googlers about how they can address the ethical issues that arise while at work. In the future, Google is planning to make this course accessible for everyone across the company. Introducing AI Ethics Speaker Series: This series features external experts across different countries, regions, and professional disciplines. So far, eight sessions have been conducted with 11 speakers covering topics from bias in natural language processing (NLP) to the use of AI in criminal justice. AI fairness: A new module on fairness is added to Google’s free Machine Learning Crash Course. This course is available in 11 languages and is being used by more than 21,000 Google employees. The fairness module explores different types of human biases that can corp in the training data and also provide strategies to identify and evaluate their effects. Review structure for new projects, products, and deals Google has employed a formal review structure to check the scaling, severity, and likelihood of best- and worst-case scenarios of new projects, products, and deals. This review structure consists of three core groups: Innovation team: This team consists of user researchers, social scientists, ethicists, human rights specialists, policy and privacy advisors, and legal experts. They are responsible for day-to-day operations and initial assessments. Senior experts: This group consists of senior experts from a range of disciplines across Alphabet Inc.. They provide technological, functional, and application expertise. Council of senior executives: This group handles the decisions that affect multiple products and technologies. Currently, more than 100 reviews have been done under this formal review structure. In the future, Google plans to create an external advisory group, which will comprise of experts from a variety of disciplines. To read more about Google’s initiatives towards ethical AI, check out their official announcement. Google won’t sell its facial recognition technology until questions around tech and policy are sorted Google expands its machine learning hardware portfolio with Cloud TPU Pods (alpha) to effectively train and deploy TensorFlow machine learning models on GCP Google kills another product: Fusion tables