Tech News - Data

Redis Labs, the provider of Redis Enterprise, announced details about its 14th consecutive quarter of double-digit growth and its annual growth of more than 60% in the company’s 2019 fiscal year. It was just last week, when Redis Labs, a California-based computer software startup, announced that it has raised $60 million in Series E financing round led by a new and leading private equity firm, Francisco Partners. Moreover, Redis Labs finished the 2019 fiscal year with more than 250 full-time employees, with its global headcount increasing 50 percent in the past year. Also, since the company scales its global go-to-market team, new offices have been opened in Austin, Texas, and Bangalore (India) to drive adoption of Redis Enterprise. Based on the record growth results and recently secured funding, Redis Labs aims to accelerate its plans in the new fiscal year across different departments including sales, marketing, and product development. This would help them meet the demands for a multi-model database that is capable of delivering the performance, deployment flexibility, and seamless scaling to advance instant experiences. Redis Labs is already continuing to expand the business with other Global 1000 enterprises such as Alliance Data, ANZ Bank, Applied Materials, Carrefour, Dick's Sporting Goods, Thomas Cook, Mercedes Benz, Nordea, UIPath, and WestPac. Other than that, Alvin Richards has been promoted to Chief Product Officer from Chief Education Officer, that he had been appointed as back in 2017. This will help him continue the company’s market leadership and deliver innovation for the multi-model database market. Redis was named as 2019 technology of the year for the second time by IDG's InfoWorld. Also, Redis secured the place of the seventh most popular database in DB-Engines ranking among more than 300 databases. Apart from having the highest rating among the top seven database providers, it is also the first database that achieved 1 billion launches on Docker Hub in 2018. “Redis Enterprise delivers the requirements for a multi-model cloud-native database that operates at record-breaking performance with unmatched cost efficiency”, mentioned Ofer Bengal, co-founder, and CEO at Redis Labs in an email sent to us. RedisGraph v1.0 released, benchmarking proves its 6-600 times faster than existing graph databases Redis Cluster Features Overview Redis Labs moves from Apache2 modified with Commons Clause to Redis Source Available License (RSAL)

Just two days ago, Perfecto, a Perforce Software company introduced Perfecto Codeless, an AI-driven codeless testing solution for eliminating the need for coding skills in the dev testing process. Perfecto Codeless will help the development teams to automate the process of writing test scripts as it comes with machine learning (ML) capabilities. The scripts will be allowed to run continuously and fix themselves without disrupting operations. Eran Yaniv, Founder, and CEO at Perfecto wrote to us in an email, “Across our customer base, the number one cause of automation failure is scripting issues. This a huge barrier for achieving good test automation and teams making their way towards continuous testing. With the introduction of Perfecto Codeless, we are harnessing the power of machine learning to offer the next generation of codeless automation testing capabilities. By eliminating the need to write and maintain test scripts, teams save time and can focus on more complex tasks.” The development teams have the coding skills for writing Selenium or Appium scripts but it is better if their time is spent on product development and innovation. Perfecto Codeless comes with tools that help the teams to quickly generate quality test scripts and maintain them. Features of Perfecto Codeless Smart capabilities to maintain scripts Perfecto’s ML capabilities address object maintenance issues within the code. If there is a need for deleting, moving or changing the code, Perfecto Codeless makes it happen agnostically without delaying the process. Interconnected components All the components of the testing process are connected right from creation to execution and analysis. Codeless automation in the cloud Perfecto Codeless provides codeless test automation in the cloud that allows teams to manage the pace and demands that come with test automation. Perfecto Codeless also provides the flexibility, performance, and scalability needed to ensure quality throughout the SDLC. Eran Kinsbruner, Chief Evangelist at Perfecto wrote to us in an email, “In recent years, codeless test automation has become a top priority for testers, as well as the developers that aim to expedite their test creation and maximize testing reliability. These professionals are looking at codeless as a preferred solution to embed into their testing responsibilities. Perfecto Codeless will take DevOps to the next level, relieving testers and developers of the time-intensive responsibility of coding and giving them time back to focus on product development and innovation to help accelerate the software delivery lifecycle (SDLC) for their business.” Rachel Batish’s 3 tips to build your own interactive conversational app DeepMind researchers provide theoretical analysis on recommender system, ‘echo chamber’ and ‘filter bubble effect’ Waymo to sell its 3D perimeter LIDAR sensors to companies outside of self-driving  

MLflow 0.8.0 released with improved UI experience and better support for deployment Last week, the team at Databricks released MLflow 0.8.0. MLflow, an open source platform used for managing end-to-end machine learning lifecycle. It is used for tracking experiments and managing and deploying models from a variety of ML libraries. It is also responsible for packaging ML code in a reusable and reproducible form in order to share the same with other data scientists. MLflow 0.8.0 features In MLflow 0.8.0, the SageMaker and pyfunc server support the ‘split’ JSON format, which helps the client to specify the order of columns. With MLflow 0.8.0, the server can now pass the gunicorn option. This is because as gunicorn uses threads instead of processes and saves memory space. This version also brings in TensorFlow 1.12 support. With this version, there’s no need of loading Keras module at predict time. Major change In MLflow 0.8.0 version, [CLI] mlflow sklearn server has been removed in favor of mlflow pyfunc serve, as it takes the same arguments but works against any pyfunc model. Major improvements in MLflow 0.8.0 This version includes various new features including improved UI experience and support for deploying models directly to the Azure Machine Learning Service Workspace. Improved MLflow UI Experience In this version, the metrics and parameters are by default grouped together in a single tabular column in order to avoid an explosion of columns. The users can customize their view by sorting the parameters and metrics. They can click on each parameter or metric in order to view them in a separate column. This makes the user experience better. The runs which are nested inside other runs can now be grouped by their parent-run. They can also be expanded or collapsed altogether. By calling mlflow.start_run or mlflow.run, a run can be nested. Though MLflow gives each run a UUID by default, one can also now assign a name to a run and also can edit the names. It makes the process easy as it is easier to remember the name than a number. There’s no need to reconfigure the view each time one uses it, as the MLflow UI remembers the filters, sorting and column setup done in browser local storage. Support for Deployment of models to Azure ML Service In this version, the Microsoft Azure Machine Learning deployment tool has been modified for deploying MLflow models packaged as Docker containers. One can use the mlflow.azureml module to package a python_function model into an Azure ML container image. Further, this image can be deployed to the Azure Kubernetes Service (AKS) and the Azure Container Instances (ACI) platforms. Major bug fixes The server works better in this version even when the environment and run files are corrupted. The Azure Blob Storage artifact repo now supports Windows paths. In the previous version, deleting the default experiment caused recreation of the same. But with MLflow 0.8.0 this problem has been fixed. Read more about this news on Databricks’ blog. Introducing EuclidesDB, a multi-model machine learning feature database Google releases Magenta studio beta, an open source python machine learning library for music artists Technical and hidden debts in machine learning – Google engineers’ give their perspective

Researchers from SenseTime Research and Nanyang Technological University have broken the record to train ImageNet/AlexNet in 1.5 minutes. The previous record was held by a model developed by the researchers at TenCent, a Chinese tech giant, and Hong Kong Baptist University, that took four minutes. This is a significant 2.6 times speedup over the previous record. The SenseTime and Nanyang team used a communication backend called “GradientFlow” along with a set of network optimization techniques to reduce the deep neural network (DNN) model training time. The researchers also proposed a technique called “lazy allreduce” to combine multiple communication operations into a single one. The researchers say that high communication overhead is one of the major performance bottlenecks for distributed DNN training across multiple GPUs. To combat this issue, one of the techniques used was increasing the batch size and running through the dataset quickly to process more samples per iteration. They also used a mixture of half-precision floating point, aka FP16, as well as single-precision floating point, FP32. Both these techniques reduce the memory bandwidth pressure on the GPUs used to accelerate the machine-learning math in hardware, but cause some loss of accuracy. How does GradientFlow work? GradientFlow is a software toolkit, to tackle the high communication cost of distributed DNN training. It is a communication backend that sla shed training times on GPUs, as described in their paper, published earlier this month. GradientFlow employs lazy allreduce, to reduce network cost, and improves network throughput by fusing multiple allreduce operations into a single one. It employs “coarse-graining sparse communication” to reduce network traffic and sends only important gradient chucks. Every GPU stores batches of data from ImageNet and uses gradient descent to crunch through their pixels. These gradient values are passed onto server nodes in order to update the parameters in the overall model. This is done using a type of parallel-processing algorithm known as allreduce. Trying to ingest these values, or tensors, from hundreds of GPUs at a time will result into bottlenecks. GradientFlow increases the efficiency of the code by allowing the GPUs to communicate and exchange gradients locally before final values are sent to the model. “Instead of immediately transmitting generated gradients with allreduce, GradientFlow tries to fuse multiple sequential communication operations into a single one, avoiding sending a huge number of small tensors via network,” the researchers wrote. Lazy allreduce Lazy allreduce fuses multiple allreduce operations into a single operation with minimal GPU memory copy overhead. On completing a backward computation, a layer with learnable parameters generates one or more gradient tensors. Every tensor is allocated a separate GPU memory space by the baseline system. With the help of lazy allreduce, all gradient tensors are placed in a memory pool. Lazy allreduce waits for the lower layer’s gradient tensors, until the total size of waited tensors  is greater than a given threshold θ. Then, a single allreduce operation is performed on all waited gradient tensors. This avoids transmitting small tensors via network and improves network utilization. Coarse-grained sparse communication (CSC) To further reduce network traffic with high bandwidth utilization, the researchers have proposed coarse-grained sparse communication to select important gradient chunks for allreduce. The generated tensors are placed in a memory pool with continuance address space, based on their generated order. The CSC will equally partition the gradient memory pool into chunks. Each chunk contains a number of gradients. In this research, each chunk contains 32K gradients and the CSC partitions the gradient memory pool of AlexNet and ResNet-50 into 1903 and 797 chunks respectively. A percent (e.g., 10%) of gradient chunks are selected as important chunks at the end of each iteration. Design of coarse-grained sparse communication (CSC) Conclusion GradientFlow improves network performance for distributed DNN training. When training ImageNet/AlexNet on 512 GPUs, the researchers achieved up to 410.2 speedup ratio, and completed 95-epoch training in 1.5 minutes, outperforming existing approaches. You can head over to the research paper for a more in-depth performance analysis of the model proposed. Generating automated image captions using NLP and computer vision [Tutorial] Facebook’s artificial intelligence research team, FAIR, turns five. But what are its biggest accomplishments? Exploring Deep Learning Architectures [Tutorial]  

Blue Oak Council Inc, a Delaware nonprofit corporation, published a model license version 1.0.0, last week. The new license demonstrates all the techniques used by the licenses to make the software free and simple for everyone to use and build on. The licensing materials published by Blue Oak is in everyday language, making it easy for developers, lawyers, and others to understand software licensing without relying on legal help. Blue Oak model license 1.0.0 comes with information regarding purpose, acceptance, copyright, notices, excuse, patent, reliability, and no liability. The license states that it provides everyone with as much permission to work with the software as possible. It also protects the contributors from liability. It states that users must agree to the rules of the license to receive it. Users should refrain from doing things that would defy the rules of the license. Additionally, it states that everyone who gets a copy of any part of the software (with or without changes) also receives a text of this license or a link to Blue Oak Council license 1.0.0. Also, in case anyone notifies the users in writing that they have not complied with Notices, then they can keep their license by taking all the practical and necessary steps to needed to comply within 30 days, post-notice. If users fail to follow this, the license will end immediately. Apart from this, Blue Oak Council has also published certain example provisions for contracts and grants, along with a corporate open source policy that helps with the permissive licenses. There’s also a list of permissive public software licenses on the OSI and SPDX lists. These licenses have been rated from gold to lead, based on criteria such as the clarity of drafting, simplicity, and practicality of conditions. For more information, check out the official Blue Oak Council blog post. Red Hat drops MongoDB over concerns related to its Server Side Public License (SSPL) Neo4j Enterprise Edition is now available under a commercial license Free Software Foundation updates its licensing materials, adds Commons Clause and Fraunhofer FDK AAC license

Microsoft released the version 2.6 of their popular deep learning framework, CNTK or Microsoft Cognitive Toolkit, last week. CNTK v2.6 explores features such as an added .NET support, efficient group convolution, improved sequential convolution, more operators, and ONNX feature update among others. Added .Net Support The Cntk.Core.Managed library has now been converted to .Net Standard. It now supports .Net Core and .Net Framework applications on both Windows as well as Linux. .Net developers will now be able to restore CNTK Nuget packages. To restore the CNTK Nuget packages, use the new .Net SDK style project file with package management format set to PackageReference. Efficient group convolution With CNTK v2.6, the group convolution has been updated. The updated implementation uses cuDNN7 and MKL2017 APIs directly instead of having to create a sub-graph for group convolution (slicing and splicing). This leads to improved experience in terms of both performance and model size. Improved Sequential Convolution Sequential Convolution implementation has also been updated in CNTK v2.6. The new implementation creates a separate sequential convolution layer. This layer offers support for broader cases, such as, where stride > 1 for the sequence axis. So, if sequential convolution is performed over a batch of one-channel black-and-white images then these images will have the same fixed height of 640 with the width of variable lengths. The width is then represented by the sequential axis. More Operators More support has been added in CNTK v2.6 for operators such as depth_to_space and space_to_depth, Tan and Atan, ELU, and Convolution. depth_to_space and space_to_depth There are breaking changes in the depth_to_space and space_to_depth operators. These two operators are updated to match the ONNX specification. Tan and Atan Support has been added for trigonometric ops Tan and Atan. ELU Support added for alpha attribute in ELU op. Convolution The auto padding algorithms of Convolution have been updated to produce symmetric padding at best effort on CPU, without influencing the final convolution output values. This leads to an increase in the range of cases which could be covered by MKL API and also improves the performance, E.g. ResNet50. Updated ONNX CNTK's ONNX import/export has been updated to support ONNX 1.2. A major update has been added on how the batch and sequence axes are handled in export and import.  CNTK's ONNX BatchNormalization op export/import has been updated to the latest spec. A model domain has been added to the ONNX model export. Support has also been added for exporting alpha attribute in ELU ONNX op. Change in Default arguments order There is a major updated to the arguments property in CNTK python API. The default behavior has been updated so now it returns the arguments in python order instead of in C++ order. Because of this, it will return arguments in the same order as they are fed into ops. Bug Fixes Improved input validation added for group convolution. Validation added for padding channel axis in convolution. Proper initialization added for ONNX TypeStrToProtoMap. The Min/Max import implementation has been updated to handle variadic inputs. There are even more updates that come with CNTK v2.6. For more information on those, check out the CNTK official release notes. The Deep Learning Framework Showdown: TensorFlow vs CNTK Deep Learning with Microsoft CNTK ONNX 1.3 is here with experimental function concept

Yesterday, Tyler Tringas, co-founder of Earnest Capital went live on Hacker News to answer questions. For those not aware, Earnest Capital provides funding and mentorship for bootstrappers, indie startups, and hackers mostly in SaaS, e-commerce, and scalable online education. Earnest Capital has been receiving a lot of attention because of their different investing structure than traditional VCs or accelerators. They have a novel Shared Earnings Agreement investing model. Key attributes of a Shared Earnings Agreement by Earnest Capital We invest upfront capital at the early-stage of businesses. Typically (but not always) after a product has launched, but before the founders go full-time. We agree on a Return Cap which is a multiple of the initial investment (typically 3-5x) We don’t have any equity or control over the business. No board seats either. You run your business as you see fit. As your business grows we calculate what we call “Founder Earnings” and Earnest is paid a percentage. Essentially we get paid when you and your co-founder get paid. Founder Earnings = Net Income + any amount of founders’ salaries over a certain threshold. If you want to eat ramen, pay yourselves a small salary, and reinvest every dollar into growth, we don’t get a penny and that’s okay. We get earnings when you do. Unlike traditional equity, our share of earnings is not perpetual. Once we hit the Return Cap, payments to Earnest end. In most cases, we’ll agree on a long-term residual stake for Earnest if you ever sell the company or raise more financing. We want to be on your team for the long-term, but don’t want to provide any pressure to “exit.” If you decide you want to raise VC or other forms of financing, or you get an amazing offer to sell the company, that’s totally fine. The SEA includes provisions for our investment to convert to equity alongside the new investors or acquirers. The Hacker News conversation was a big hit, with Tyler mostly answering questions and offering pieces of advice to upcoming entrepreneurs while also explaining their novel strategy. Per Tyler, Earnest Capital works like a “profit-share + a SAFE”. The primary function is for them is to share in the profit (or more specifically "founder earnings") of the business alongside the founder(s). If later a business owner decides to sell the business or raise a big equity round later, Earnest converts into a SAFE. On how it is different from a Venture Capitalist Investment A user asked, “If I read your agreement correctly, your terms are so that you invest $150k in what is or nearly is a bootstrapped business, on what essentially seems like a profit share basis, and expect to get paid for your doing so until you've made $3M?” Tyler’s response, “$3m?! No. We have a Return Cap which is negotiated on a per deal basis but we guide toward 3-5x the initial investment. This post walks through each of the terms in detail.” Traditional VC model usually works in cycles. The founders raise some money, then they build and sell, raise some more, build and sell. This cycle continues by the time they have raised enough money to be at least close to a profit. “How does that work with bootstrappers? Ideally, they'd only have to raise money once (from you), but what happens after the $100k (example) run out and the business is only generating, let's say, $2k/month? Back to 9-to-5?” asked another user. Tyler argued that this is true only in some cases. Earnest Capital’s goal is for founders to get to “personal break-even, where they can pay themselves enough to work on the business full-time, by the time our investment runs out. Some percentage of these will fail (startups are hard) and we're expecting that”, he added. On comparison with TinySeed People also appreciated the novel non-VC funding space asking Tyler to do a quick compare/contrast with TinySeed, (TinySeed is also a startup accelerator designed for bootstrappers) of similar ilk and recency. Tyler commented, “Specific to the funding model, we both do a kind of profit-share, with the main difference being that Earnest's repayment will usually happen earlier (assuming the business is successful) but is capped, Tinyseed payments would be smaller in the earlier years and keep growing over time perpetually. Neither one is "better" and I probably wouldn't advise founders to choose between an offer from both on the basis of just the funding model.” On their 3-5x return cap People also had concerns regarding the 3-5x return cap. Some called Earnest Capital “more of a charity than a profit-making enterprise?” to the extent of calling it altruistic. A Hacker news user observed, “For the math to work out, with a 3x cap on earnings, you need 33% of the businesses to be successful just to get your money back. At 5x you still need 20%. And that is over however long it takes for those companies to reach payback, which could be measured in decades for some of the companies.” To this, Tyler said, “We also have a residual, uncapped, % option if the founder ever sells the business. This keeps us aligned with the founders to keep helping them grow the value of the business for the long-term even after the Return Cap is paid back.” He added that they are preparing to fine tune the return cap model building, measuring, learning, and iterating as we go. Basically, he added, “By default, we don't take equity (shares, a board seat, none of that). If you decide to raise a round of equity financing (ie VC) we could convert into equity alongside them and if you sell the company we get a % of that.” A user countered it saying that “This return cap is stated as 9.5% in the spreadsheet. But, where did that come from? Is this is a stock option, convertible note, or equity position?” Another user added, “They don't explain it, but it sounds like the whole deal is effectively seed financing where they eventually get a 9.5% stake, but also with a 3-5x loan interest payment once you make money (which they're framing as "Shared Earnings"). And they're trying to hide the 9.5% part.” Tyler offered no comments on this thread. On being asked why return cap is better than just taking out like a 20-30% APR business loan, Tyler said, “At the risk of not answering the question, I'd say no form of capital is "better" than any other. Capital is a tool and the job of the founder is to find the option (both on payments term and other aspects like mentorship or personal exposure) that best aligns with their goals.” On his thoughts on Jerry Neumann’s theory Jerry Neumann, who is a Venture Capitalist at Neu Venture Capital has a theory about why "there's a reason why for decades, there were only bank loans and VC and not much in-between." Per his theory, there are 3 categories of companies (determined by the alpha value of the power-law distribution they're in): Companies where the risk and the upside potential are small. This is where bank loans are focused. Companies where the risk is enormous but the upside potential is "meh". Companies where the risk is enormous but the upside potential is also enormous. This is where VC is focused, and it's why they're all about finding those few big hits because this covers all the losses (or mediocre performance) of the rest. Neumann appears pretty confident about this hypothesis; not because he can explain the underlying phenomenon, but simply because until now he's not seen much successful funding for companies that's neither VC nor bank loans. A hacker news user points, that “if his hypothesis is right, then Earnest Capital is targeting companies of type 2: investments with enormous risk (comparable to that of a high-growth startup) but at the same time you're hard-capping your upside at 5x. That seems madness.” Tyler comments, “I like Jerry's work a lot but come to a different conclusion. My basic thesis is that we're in the deployment age of the internet/web/mobile era and there is a whole new wave of a lot lower risk and a bit fewer reward opportunities for companies to bring the "peace dividend" of the software areas into markets that are not winner-take-all. The upside is these businesses are much more capital efficient, can scale and potentially produce much higher returns than SMBs from previous eras. The downside is they have no collateral and are thus completely unbankable for traditional small business lending. We need a new default form of capital for entrepreneurs and we are trying to build it.” Overall people were generally appreciative of Earnest Capital and wished the company success. Here are some positive responses. “Hi Tyler, this is amazing! Going through the FAQ it seems, you're not investing in India right now! Would love to know if and when you do!” “Very cool, Tyler. Just applied. :)” “Congratulations on launching! Sounds like a nice compliment to IndieVC and TinySeed. I'm glad to see innovation here, interesting times!” “Awesome to see innovation from capital providers on the instrument in the wild. As a niche market founder wish option like Earnest existed when we were raising early financing.” We recommend you to go through the entire thread on Hacker News. It makes for a very insightful conversation. A Quick look at ML in algorithmic trading strategies Why Google kills its own products Mary Meeker, one of the premier Silicon Valley investors, quits Kleiner Perkins to start her own firm.

Last year, the US Commerce Department issued a notice named advance notice of proposed rulemaking (ANPRM) that lists some emerging technologies on which export controls will be employed. The list includes fourteen categories including AI, quantum computing, robotics, advanced materials, and advanced surveillance technologies, among others. This notice called for public opinion on the criteria according to which these emerging technologies will be identified that are essential to U.S. national security. The public was requested to submit their comments on or before December 19, 2018, which is now extended to January 10, 2019. After identifying an emerging or foundational technology based on public comments, the Commerce Department will be authorized under the Export Control Reform Act (ECRA) of 2018 to establish "appropriate controls" on "emerging and foundational technologies". Will employing export rules on emerging technologies be successful? A technology policy researcher at the Massachusetts Institute of Technology, R. David Edelman believes that it is nearly impossible to classify technologies in two categories: military or commercial. He told to the New York Times, “trying to draw a line between what is military and what is commercial is exceedingly difficult. It may be impossible.” Another point to note here is that the research on these ever dynamic technologies is often done by scientists and engineers collaboratively all over the world, so we cannot really claim that the product is entirely developed by America. Also, companies and other researchers working on these technologies open source their work in hopes that some other researcher will be able to further develop the tool or technology. This is why policy experts believe that these US restrictions will have very little effect on the progress of AI in China and other countries. The government is unlikely to restrict companies and universities from publishing results of their AI research. But Greg Jaeger, a lawyer at the law firm Stroock & Stroock & Lavan who deals with export controls told NYT that the government could restrict foreign access to that information. One of the Hacker News readers wondered, “I'm curious how export restrictions would affect open source projects like Tensorflow and PyTorch. Would they be forced to become closed source? Could the license just include a disclaimer: "You're not allowed to use this if you're in one of the following countries: ..."? Would sites like Gitlab and Github be forced to implement per-repo geoblocking? Could they somehow be moved to ownership by a non-American entity that wasn't subject to such code? Does a US citizen contributing to a non-US open source ML project constitute a breach of export controls?”. They could also put control over the export of cloud-computing technology and computer chips related to artificial intelligence. These restrictive rules can prevent researchers from other countries from working on certain technologies in the US. They would rather choose other countries such as Europe. “It might be easier for people to just do this stuff in Europe,” said Jason Waite, a lawyer with the firm Alston & Bird who specializes in international trade, in the NYT interview. Patreon speaks out against the protests over its banning Sargon of Akkad for violating its rules on hate speech Mozilla v. FCC: Mozilla challenges FCC’s elimination of net neutrality protection rules Twitter prepares for mid-term US elections, with stronger rules and enforcement approach to fight against fake accounts and other malpractices

This is a big win for the Artificial Intelligence community. IEEE has released a  statement from the IEEE Board of Directors stating that the committee will now support the inclusion of ethical considerations in the design and deployment of autonomous and intelligent systems (A/IS). The IEEE committee recognizes that present AI systems present new social, legal and ethical challenges. They also have to address issues of systemic risk, diminishing trust, privacy challenges and issues of data transparency, ownership and agency. Therefore, there is a need for developers of such systems to use practices and standards that respect and acknowledge the ethical obligation of such systems in their human and social context. Concrete steps taken by IEEE A/IS systems should be developed and operated in a manner that respects internationally recognized human rights. A/IS developers should consider impact on individual and societal well-being to be central in development. Developers should respect each individual’s ability to maintain appropriate control over their personal data and identifying information. Developers and operators should consider the effectiveness and fitness of A/IS technologies for the purpose of their systems. Technical basis of particular decisions made by an A/IS should be discoverable. A/IS should be designed and operated in a manner that permits production of an unambiguous rationale for the decisions made by the system. Designers of A/IS creators should consider and guard against potential misuses and operational risks. Designers of A/IS should specify and operators should possess the knowledge and skills required for safe and effective operation. To that extent, the IEEE committee has taken various initiatives to build ethically aligned AI systems. In March, they released a report, “Ethically Aligned Design – A Vision for Prioritizing Human Well-being with Autonomous and Intelligent Systems, Edition 1.0,” that sets forth scientific analysis and resources, high-level principles and actionable recommendations for ethical implementation of A/IS. They also launched the IEEE Tech Ethics program which seeks to ensure that ethical and societal implications of technology become an integral part of the development process by driving conversation and debate on these issues. The IEEE Code of Ethics also showcases IEEE’s commitment to ethical design and the societal implications of intelligent systems. In a statement the IEEE committee said, “IEEE is committed to developing trust in technologies through transparency, technical community building, and partnership across regions and nations, as a service to humanity. Measures that ensure that A/IS are developed and deployed with appropriate ethical consideration for human and societal values will enhance trust in these technologies, which in turn will increase the ability of the technologies to achieve much broader beneficial societal impacts.” The news was quite well received by the developer community after John C. Havens, Executive Director at The IEEE Global Initiative on Ethics of Autonomous and Intelligent Systems shared the news on Twitter. Users called it as arguably the most globally impactful step in this space and a milestone for all. https://twitter.com/jameshorton/status/1145900183042973698 https://twitter.com/GReal1111/status/1145826945336262662   Some pointed out that all tech companies should sign on to this statement. https://twitter.com/Dktr_Sus/status/1145866352176979968 Read the full report here. The US puts Huawei on BIS List forcing IEEE to ban Huawei employees from peer-reviewing or editing research papers. IEEE Standards Association releases ethics guidelines for automation and intelligent systems IEEE Computer Society predicts top ten tech trends for 2019: assisted transportation, chatbots, and deep learning accelerators among others.

This year, in the FIFA World Cup researchers and scientists, tried to use Artificial Intelligence (AI) to predict the outcomes of all 64 matches. That did not work out well, accounting to probability and human nature that cannot be predicted. Now we see another implementation of AI in a major global sports event. This time it’s not to predict outcomes, but to identify players with facial recognition. In 2020, facial recognition will be used for the first time widely in an Olympic event to identify athletes. The Japanese IT firm NEC Corp will provide the facial recognition system. The facial recognition system will also be used in the 2020 Paralympics;  it was tested in the Rio 2016 Olympics. The system is built around an AI engine called NeoFace. In addition to athletes, the system will be used to identify volunteers, media, and other staff. It will be used to identify around 300,000 people across more than 40 venues. The Olympics will begin on July 24, 2020. People attending the event are expected to submit their data in advance before the Olympics start. It will be approved and stored in a database and used to identify players before entry. The system will link the person’s facial data with an IC card that will be carried by them. So entry would be permitted only if the facial data stored in the database matches the data stored in the IC carried by the person. Tokyo 2020 has security challenges since venues are not that large. This will result in long wait times before the players can get into the venues. With the summer heat in Tokyo, this presents a problem for the players. The events will be spread out across the metropolitan area, and people will have to authenticate themselves at every entry. The NeoFace system is introduced to address these problems in the Tokyo Olympics venues. The facial recognition system is aimed at strengthening security and minimizing waiting times for athletes. NeoFace will also help with identifying forged ID cards and help athletes avoid the stress of waiting in long lines for identification. NEC has substantial experience in the facial recognition field and their technology has been used at airports for several years. The Tokyo Olympics 2020 may be the Olympics event with most security implemented yet. For more information, you can check out the coverage by Reuters. Read Next: Microsoft’s Brad Smith calls for facial recognition technology to be regulated Amazon is selling facial recognition technology to police Admiring the many faces of Facial Recognition with Deep Learning

New tool Olympus, SAP's collaboration with Red Hat, blockchain-powered Visa B2B Connect, and Dialogflow Enterprise Edition in today's top stories around data science news. An instant REST API for any AI model Olympus – A new tool that instantly creates a REST API for any AI model Olympus is a command-line tool to deploy a pre-trained machine learning or deep learning model as a REST API, in seconds, thus eliminating the need for developers to manually create the REST APIs. “One of the key differences between Olympus and TensorFlow Serving is that, while TF Serving is optimized for the production environment, Olympus is currently more geared towards the development phase,” Olympus developers announced. To install Olympus, run the code pip install olympus. Big data meets containerization SAP Vora introduced into Red Hat OpenShift Container Platform SAP Vora solution on Red Hat OpenShift Container Platform is an integrated solution that pairs enterprise-grade Kubernetes with actionable big data insights. Key features of the integrated offering include: On-demand in-memory big data analytics, easier management of big data analytics at scale, easier integration of SAP Vora with SAP HANA, and better support for agile development around big data use cases. Visa B2B for cross-border corporate payments Visa kicks off pilot phase of “Visa B2B Connect” blockchain-based platform, commercial launch in mid-2018 Visa has announced the pilot phase of its blockchain-based platform Visa B2B Connect. The credit card giant had previewed the global payment platform in October 2016. Using blockchain-based architecture, Visa B2B Connect simplifies existing cross-border corporate payments by sending transactions over Visa’s network from the bank of origin directly to the recipient bank. Following the first phase, the commercial launch of the platform is planned for mid-2018. Google Dialogflow to power conversational interactions Google rolls out paid enterprise edition of Dialogflow with added speech integration Google has announced the beta release of enterprise edition of Dialogflow, its tool for building chatbots and other conversational applications. The enterprise edition offers greater flexibility and support for large-scale businesses, and has built-in support for speech recognition,  enabling developers to build rich voice-based applications. The enterprise Edition offers unlimited pay-as-you-go voice support. Companies like Uniqlo, Policybazaar.com and Strayer University have already used Dialogflow to design and deploy conversational experiences, Google said. Dialogflow was formerly known as API.AI, until its acquisition by Google last year.

BBC is set to embrace the AI revolution with open arms. Last week, it announced a new BBC 4.1 AI TV. This aims to bring “two nights of experimental programming” featuring the new and classic programmes that explore AI. The programme will launch on BBC four on two nights: 4-5 September with Dr. Hannah Fry as the presenter. It will also feature a “virtual co-presenter". BBC 4.1 AI TV promo ‘BBC 4.1 AI TV’ will feature 'Made by Machine: When AI met The Archive', an experimental programme partly made by artificial intelligence, trained to show information dating back to 1953 from well over 250,000 TV programmes. This approach manually would have been impractical as it would take hundreds of hours. But, with the help of latest AI technology from BBC Research & Development, it provided BBC four with a more manageable selection of shows. “The AI learnt what BBC Four audiences might like, based on the channel’s previous schedules and programme attributes, and then ranked programmes it thought were most relevant,” says BBC. It will be broadcasting a selection of programmes that haven’t been seen in years. The programme on BBC 4.1 AI TV features four sections of archive clips edited together that follows the sequence as mentioned below: In the first segment, the AI learns to detect different attributes of the scene such as what a scene consists of, the type of landscape, the objects present, whether people are featured and people’s apparel. This helps the people learn about how a compilation is created with each scene following up from the last. For the second segment, the subtitles or archive programmes are scanned to put together a footage by looking for links between words, topics and themes. The third segment consists of AI analyzing the activity levels on screen ( whether they are high or not ). It then attempts to create a compilation that moves back and forth between high energy and low energy scenes. The fourth sequence combines all its learned to create an altogether new piece of content. According to Cassian Harrison, Channel Editor, BBC Four, “ In collaboration with the BBC's world-beating R&D department, AI TV will explore -- and demonstrate just how AI and machine learning might inform and influence programme-making and scheduling, while also resurfacing some gems from the BBC Four archive along the way”. Also, “Helping BBC Four scour BBC’s vast archives more efficiently is exactly why we’re developing this kind of AI --  and has massive benefits for BBC programme makers and audiences -- Made By Machine: When AI Met The Archive gives people an unprecedented look under the hood” says George Wright, Head of Internet Research and Future Services, BBC R&D. For more coverage on this news, check out the official BBC announcement. Baidu announces ClariNet, a neural network for text-to-speech synthesis Nvidia and AI researchers create AI agent Noise2Noise that can denoise images How Amazon is reinventing Speech Recognition and Machine Translation with AI  

Google's AIY Vision Kit, Amazon's Alexa for Business, and more in today's top stories in data science news. Amazon Web Services in news Amazon is putting "Alexa for business" Amazon Web Services has announced a new initiative to get companies use Alexa in the office. Under the plan, the virtual assistant will help employees launch conference calls, organize room bookings, and even discuss their expenses. With the new scheme, Alexa for Business, companies will be given the tools to manage a fleet of Alexa-enabled devices. Admins will be able to enroll users, enable and disable skills, and connect Alexa to their conferencing equipment. They’ll also be able to build their own apps for the assistant, with Amazon suggesting functions like helping with directions around the office, reporting problems with equipment, and ordering new supplies. Users will also be able to access their company’s apps from home devices, checking what’s on their office calendar and remotely joining meetings. Alexa for Business is being seen as a direct competition to other virtual assistants like Apple Siri, Google Assistant, or Microsoft Cortana. DigitalGlobe to leverage AWS suite of machine learning capabilities DigitalGlobe has migrated its entire 100-petabyte imagery library to Amazon Web Services, thus giving its customers instant access to Amazon's vast library of geospatial images. DigitalGlobe's sister division Radiant Solutions and its partner ecosystem are also leveraging AWS’s frameworks and tools to build machine learning applications that allow their customers to incorporate valuable geospatial information extracted from commercial satellite imagery into their workflows. “Few companies work with the sheer volume of data that DigitalGlobe does. When working at this volume, it’s nearly impossible to scale and rapidly innovate without the cloud,” said AWS VP Teresa Carlson said, adding that DigitalGlobe was the first customer to use AWS Snowmobile – AWS’s Exabyte-scale data transfer service that uses a 45-foot long ruggedized shipping container pulled by a semi-trailer truck – to move their massive image library to AWS. Google in news Introducing the AIY Vision Kit: Add computer vision to your maker projects Google's AIY Team has announced its next project: the AIY Vision Kit — an affordable, hackable, intelligent camera. The AIY Vision Kit is easy to assemble and connects to a Raspberry Pi computer. Based on user feedback, this new kit is designed to work with the smaller Raspberry Pi Zero W computer and runs its vision algorithms on-device so there's no cloud connection required. The kit materials list includes a VisionBonnet, a cardboard outer shell, an RGB arcade-style button, a piezo speaker, a macro/wide lens kit, flex cables, standoffs, a tripod mounting nut and connecting components. "For those of you who have your own models in mind, we've included the original TensorFlow code and a compiler. Take a new model you have (or train) and run it on the the Intel® Movidius™ MA2450," Google said adding that users can extend the kit to solve their real world problems. A blockchain for health data Health Wizz unveils blockchain platform to give patients control of health data Health Wizz announced the upcoming launch of its blockchain-based solution designed to address the mounting problem of electronic health records, and provide patients more power over their own health information. Using the Health Wizz platform, every patient would become the arbiter of his or her own medical records. "Each time medical records are produced – by a doctor’s appointment, ER visit, hospital intake or self-reporting app – the platform would standardize them using a specification known as Fast Healthcare Interoperability Resources," the company said. "Once done, the records are secured on the user’s own mobile devices in an encrypted space accessible only by that user’s own private cryptographic keys." To power its system, Health Wizz today announced a pre-sale of its digital token, which will run from Nov. 30 until February 2018. Proceeds will be used to develop the platform further and augment already existing venture capital investments. The formal launch of the platform will happen in March 2018. Other data science news H2O.ai secures $40 million to democratize artificial intelligence for the enterprise H2O.ai announced it has completed a $40 million Series C round of funding led by Wells Fargo and NVIDIA with participation from New York Life, Crane Venture Partners, Nexus Venture Partners and Transamerica Ventures, the corporate venture capital fund of Transamerica and Aegon Group. The Series C round brings H2O.ai's total amount of funding raised to $75 million. The new investment will be used to further democratize advanced machine learning and for global expansion and innovation of Driverless AI, an automated machine learning and pipelining platform that uses “AI to do AI.” H2O’s signature community conference, H2O World will take place on December 4-5, 2017 at the Computer History Museum in Mountain View, Calif. Impetus Technologies to host meetup on anomaly detection techniques using Apache Spark Big data company Impetus Technologies announced it will host a complimentary meetup "Anomaly Detection Techniques and Implementation Using Apache Spark" on Tuesday, December 5, 2017 from 6-8 pm Pacific time at the Larkspur Landing Hotel in Milpitas, Calif. The company said that space is limited for the event, and interested data scientists, developers and information technology (IT) professionals are asked to reserve a seat at the complimentary event by emailing at events@impetus.com. In the meetup, the StreamAnalytix team from Impetus will share insights on choosing the right anomaly detection techniques and demonstrate real-world use cases for finding variances in network traffic and financial transactions. Uptake raises $117M at $2.3B valuation for industrial predictive analytics Uptake, a SaaS startup that uses machine learning to read and understand how machines are working, and also anticipate when they may break down or need other attention, has closed a Series D round of $117 million at a post-money valuation of $2.3 billion, led by Baillie Gifford, with participation also from existing investors Revolution Growth and GreatPoint Ventures. It brings the total funding to over $250 million. “We’re on a growth trajectory now where there is virtually nothing standing in our way from being the predictive analytics market leader across every heavy industry, from oil & gas to mining and beyond,” said Uptake Co-founder and CEO Brad Keywell in a statement. CrowdRiff releases 'smart' visual content marketing platform Visual marketing software provider CrowdRiff said it has now processed over 500 million images for over 300 travel brands, and is releasing new visual marketing capabilities powered by artificial intelligence and machine learning. CEO Dan Holowack announced this new release at the DTTT Global conference in Brussels, Belgium, where he is co-presenting a session, "Making the Shift to Visual Marketing," together with Amber King, Director, U.S. Marketing at Colorado Tourism Office. "The volume of available visual content is larger than ever before, and finding the perfect visuals that meet both brand and performance goals is a time-consuming and largely manual process," Holowack said, adding that CrowdRiff's latest release addresses "the most common problems marketing teams face when producing visual content, at every stage of the visual content lifecycle."

Microsoft SQL Server Management Studio 17.6, IBM’s Deep Learning as a Service program, Intel’s nGraph, and more in today’s top stories around machine learning, deep learning, and data science news. Top Data science News Stories of the Day Azure Database Services are now generally available for MySQL and PostgreSQL. Microsoft announces the release of SSMS, SQL Server Management Studio 17.6. IBM rolls out Deep Learning as a Service (DLaaS) program for AI developers. Other Data Science News at a Glance Intel AI open sources nGraph, a framework-neutral Deep Neural Network (DNN) model compiler. With nGraph, data scientists can focus on data science rather than worrying about how to adapt their DNN models to train and run efficiently on different devices. Read more on the Intel AI Blog. Databricks introduces a new millisecond low-latency mode of streaming called continuous mode in Apache Spark 2.3, now available in Databricks Runtime 4.0 as part of Databricks Unified Analytics Platform. Read more on the Databricks Blog. IBM announces the launch of IBM Watson Data Kits, which are designed to accelerate the development of AI applications. Watson Data Kits will provide companies with pre-enriched, machine-readable, industry-specific data for building AI projects. Read more on PR Newswire. DeepL introduces DeepL API, as part of DeepL Pro. DeepL API is an interface that allows other computer programs to send texts to DeepL servers and receive high-quality translations. Read more on the DeepL Blog. The Altair 2.0 release candidate is now available, with full support for interactive Vega-Lite visualizations in Python. Features a declarative grammar of interactions for building sophisticated interactive data views from easy-to-understand building blocks. Read more on Altair Github.

PagerDuty, a cloud computing firm, had its shares jump 59% in its opening day of trading at the New York Stock Exchange, yesterday after the firm priced its first software IPO of 2019 above an elevated range. https://twitter.com/jenntejada/status/1116108997189406723 https://twitter.com/superlilia/status/1116337381559267329 PagerDuty had initially targeted a price of $19 to $21 for its IPO shares. However, PagerDuty’s shares traded $38.25 a piece, giving the company a market capitalization of around $2.8 billion.   Its stock was priced at $24 a piece, late Tuesday this week with a market cap of nearly $1.8 billion. PagerDuty had plans of selling 8.5 million shares in the IPO, but the stockholders offered about 570,000 shares. PagerDuty operates an On-Call Management platform. The company helps improve the engagement between software developers and operators as it enables them to take action in real time. It is used by developers, IT, security, and customer support domains across different companies.   The company was incorporated in 2010 and has its headquarters in San Francisco, California and its competitors include Atlassian Corp, OpsGenie, and Splunk. PagerDuty’s first financing round took place in September last year, where it was valued at $1.3 billion in a private funding round led by T. Rowe Price and Wellington Management. PagerDuty is used by over 11000 companies, extending across 350,000 employees, including Box and Slack. Its revenue hiked 48% last year, reaching $117.8 million. However, the company’s overall loss had reached $40.7 million from $38.1 million last year. Jennifer Tejada, CEO, PagerDuty, told TechCrunch about the IPO that “it is a gratifying day, especially for the co-founders who were pulling the idea together for PagerDuty...before they even launched it, and for employees who’ve been with the company for nearly as long and...turned down safer and higher-paying jobs along the way”. Tejada owns a 5.7% stake in the company and says that PagerDuty is going to serve a $25 billion market made up of about 22 million software developers.   Public reaction to the news is positive, with people congratulating PagerDuty for the win: https://twitter.com/ethankurz/status/1116340622091272194 https://twitter.com/sacca/status/1116129693617442816 https://twitter.com/alexrkonrad/status/1116351697339723779 https://twitter.com/stewart/status/1116252660905021440 Amazon stocks surge past $2000, expect Amazon to join Apple in the $1 trillion market cap Apple stocks soar just shy of $1 Trillion market cap as revenue hits $53.3 Billion in Q3 earnings 2018 Why Wall Street unfriended Facebook: Stocks fell $120 billion in market value after Q2 2018 earnings call