Tech News - Data

[box type="shadow" align="" class="" width=""]Dr. Brandon: Hello and welcome to the third episode of 'Date with Data Science'. Today we talk about decision trees in machine learning. Jon: Decisions are hard enough to make. Now you want me to grow a decision tree. Next, you'll say there are decision jungles too! Dr. Brandon: It might come as a surprise to you, Jon, but decision trees can help you make decisions easier. Imagine you are in a restaurant and you are given a menu card. A decision tree can help you decide if you want to have a burger, pizza, fries or a pie, for instance. And yes, there are decision jungles, but they are called random forests. We will talk about them another time. Jon: You know Bran, I have never been very good at making decisions. But with food, it is easy. It's ALWAYS all you can have. Dr. Brandon: Well, my mistake. Let's take another example. You go to the doctor's after your binge eating at the restaurant with stomach complaints. A decision tree can help your doctor decide if you have a problem and then to choose a treatment option based on what your symptoms are. Jon: Really!? Tell me more. Dr. Brandon: Alright. The following excerpt introduces decision trees from the book Apache Spark 2.x Machine Learning Cookbook by Siamak Amirghodsi, Meenakshi Rajendran, Broderick Hall, and Shuen Mei. To know how to implement them in Spark read this article. [/box] Decision trees are one of the oldest and more widely used methods of machine learning in commerce. What makes them popular is not only their ability to deal with more complex partitioning and segmentation (they are more flexible than linear models) but also their ability to explain how we arrived at a solution and as to "why" the outcome is predicated or classified as a class/label. A quick way to think about the decision tree algorithm is as a smart partitioning algorithm that tries to minimize a loss function (for example, L2 or least square) as it partitions the ranges to come up with a segmented space which are best-fitted decision boundaries to the data. The algorithm gets more sophisticated through the application of sampling the data and trying a combination of features to assemble a more complex ensemble model in which each learner (partial sample or feature combination) gets to vote toward the final outcome. The following figure depicts a simplified version in which a simple binary tree (stumping) is trained to classify the data into segments belonging to two different colors (for example, healthy patient/sick patient). The figure depicts a simple algorithm that just breaks the x/y feature space to one-half every time it establishes a decision boundary (hence classifying) while minimizing the number of errors (for example, a L2 least square measure): The following figure provides a corresponding tree so we can visualize the algorithm (in this case, a simple divide and conquer) against the proposed segmentation space. What makes decision tree algorithms popular is their ability to show their classification result in a language that can easily be communicated to a business user without much math: If you liked the above excerpt, please be sure to check out Apache Spark 2.0 Machine Learning Cookbook it is originally from to learn how to implement deep learning using Spark and many more useful techniques on implementing machine learning solutions with the MLlib library in Apache Spark 2.0.

Yesterday, Benjamin Groessing, a web consultant and developer at byteq, released remove.bg, a tool built on python, ruby and deep learning. This tool automatically removes the background of any image within 5 seconds. It uses various custom algorithms for the processing of the image. https://twitter.com/hammer_flo_/status/1074914463726350336 It is a free service and users don’t have to manually select the background/foreground layers to separate them. One can simply select an image and instantly download the resulting image with the background removed. Features of remove.bg Personal and professional use Remove.bg can be used by graphic designer, photographer or selfie lover for removing backgrounds. Saves time and money It saves time as it is automated and it is free of cost. 100% Automatic Apart from the image file, this release doesn’t require inputs such as selecting pixels, marking persons, etc. How does remove.bg work? https://twitter.com/begroe/status/1074645152487129088 Remove.bg uses AI technology for detecting foreground layers and separating them from the background. It uses additional algorithms for improving fine details and preventing color contamination. The AI detects persons as foreground and everything else as background. So, it only works if there is at least one person in the image. Users can upload images of any resolution but for performance reasons, the output image has been limited to 500 × 500 pixels. Privacy in remove.bg User images are uploaded through a secure SSL/TLS-encrypted connection. These images are processed and the result is temporarily stored till the time a user can download them. After which, approximately an hour later, these image files get deleted. Privacy message on the official website of remove.bg states, “We do not share your images or use them for any other purpose than removing the background and letting you download the result.” What can be expected from the next release? The next set of releases might support other kinds of images such as product images. The team at Remove.bg might also release an easy-to-use API. Users are very excited about this release and the technology used behind it. Many users are comparing it with the portrait mode on iPhone X. Though it is not that fast but users are still liking it. https://twitter.com/Baconbrix/status/1074805036264316928 https://twitter.com/hammer_flo_/status/1074914463726350336 But how strong is remove.bg with regards to privacy is a bigger question. Though the website gives a privacy note at the end but it will take more to win the user’s trust. The images uploaded to remove.bg’ cloud might be at risk. How strong is the security and what preventive measures have they taken? These are few of the questions that might bother many. To have a look at the ongoing discussion on remove.bg, check out Benjamin Groessing’s AMA twitter thread. Facebook open-sources PyText, a PyTorch based NLP modeling framework Deep Learning Indaba presents the state of Natural Language Processing in 2018 NYU and AWS introduce Deep Graph Library (DGL), a python package to build neural network graphs

Data Scientists work really hard to upskill their technical competencies. A rapidly changing technology landscape demands a continuous ramp up of skills like mastering a new programming language like R, Python, Java or something else, exploring new machine learning frameworks and libraries like TensorFlow or Keras, understanding cutting-edge algorithms like Deep Convolutional Networks and K-Means to name a few. Had they lived in Dr.Frankenstein's world, where scientists worked hard in their labs, cut-off from the rest of the world, this should have sufficed. But in the real world, data scientists use data and work with people to solve real-world problems for people. They need to learn something more, that forms a bridge between their ideas/hypotheses and the rest of the world. Something that’s more of an art than a skill these days. We’re talking about soft-skills for data scientists. Today we’ll enjoy a conversation between a father and son, as we learn some critical soft-skills for data scientists necessary to make it big in the data science world. [box type="shadow" align="" class="" width=""] One chilly evening, Tommy is sitting with his dad on their grassy backyard with the radio on, humming along to their favourite tunes. Tommy, gazing up at the sky for a while, asks his dad, “Dad, what are clouds made of?” Dad takes a sip of beer and replies, “Mostly servers, son. And tonnes of data.” Still gazing up, Tommy takes a deep breath, pondering about what his dad just said. Tommy: Tell me something, what’s the most important thing you’ve learned in your career as a Data Scientist? Dad smiles: I’m glad you asked, son. I’m going to share something important with you. Something I have learned over all these years crunching and munching data. I want you to keep this to yourself and remember it for as long as you can, okay? Tommy: Yes dad. Dad: Atta boy! Okay, the first thing you gotta do if you want to be successful, is you gotta be curious! Data is everywhere and it can tell you a lot. But if you’re not curious to explore data and tackle it from every angle, you will remain mediocre at best. Have an open mind - look at things through a kaleidoscope and challenge assumptions and presumptions. Innovation is the key to making the cut as a data scientist. Tommy nods his head approvingly. Dad, satisfied that Tommy is following along, continues. Dad: One of the most important skills a data scientist should possess is a great business acumen. Now, I know you must be wondering why one would need business acumen when all they’re doing is gathering a heap of data and making sense of it. Tommy looks straight-faced at his dad. Dad: Well, a data scientist needs to know the business like the back of their hand because unless they do, they won’t understand what the business’ strengths and weaknesses are and how data can contribute towards boosting its success. They need to understand where the business fits into the industry and what it needs to do to remain competitive. Dad’s last statement is rewarded by an energetic, affirmative nod from Tommy. Smiling, dad’s quite pleased with the response. Dad: Communication is next on the list. Without a clever tongue, a data scientist will find himself going nowhere in the tech world. Gone are the days when technical knowledge was all that was needed to sustain. A data scientist’s job is to help a business make critical, data-driven decisions. Of what use is it to the non-technical marketing or sales teams, if the data scientist can’t communicate his/her insights in a clear and effective way? A data scientist must also be a good listener to truly understand what the problem is to come up with the right solution. Tommy leans back in his chair, looking up at the sky again, thinking how he would communicate insights effectively. Dad continues: Very closely associated with communication, is the ability to present well, or as a data scientist would put it - tell tales that inspire action. Now a data scientist might have to put forward their findings before an entire board of directors, who will be extremely eager to know why they need to take a particular decision and how it will benefit the organization. Here, clear articulation, a knack for storytelling and strong convincing skills are all important for the data scientist to get the message across in the best way. Tommy quips: Like the way you convince mom to do the dishes every evening? Dad playfully punches Tommy: Hahaha, you little rascal! Tommy: Are there any more skills a data scientist needs to possess to excel at what they do? Dad: Indeed, there are! True data science is a research activity, where problems with unclear or unobvious solutions get solved. There are times when even the nature of the problem isn’t clear. A data scientist should be skilled at performing their own independent research - snooping around for information or data, gathering it and preparing it for further analysis. Many organisations look for people with strong research capabilities, before they recruit them. Tommy: What about you? Would you recruit someone without a research background? Dad: Well, personally no. But that doesn’t mean I would only hire someone if they were a PhD. Even an MSc would do, if they were able to justify their research project, and convince me that they’re capable of performing independent research. I wouldn’t hesitate to take them on board. Here’s where I want to share one of the most important skills I’ve learned in all my years. Any guesses on what it might be? Tommy: Hiring? Dad: Ummmmm… I’ll give this one to you ‘cos it’s pretty close. The actual answer is, of course, a much broader term - ‘management’. It encompasses everything from hiring the right candidates for your team to practically doing everything that a person handling a team does. Tommy: And what’s that? Dad: Well, as a senior data scientist, one would be expected to handle a team of lesser experienced data scientists, managing, mentoring and helping them achieve their goals. It’s a very important skill to hone, as you climb up the ladder. Some learn it through experience, others learn it by taking management courses. Either way, this skill is important for one to succeed in a senior role. And, that’s about all I have for now. I hope at least some of this benefits you, as you step into your first job tomorrow. Tommy smiles: Yeah dad, it’s great to have someone in the same line of work to look up to when I’m just starting out my career. I’m glad we had this conversation. Holding up an empty can, he says, “I’m out, toss me another beer, please.”[/box] Soft Skills for Data Scientists - A quick Recap In addition to keeping yourself technically relevant, to succeed as a data scientist you need to Be curious: Explore data from different angles, question the granted - assumptions & presumptions. Have strong business acumen: Know your customer, know your business, know your market. Communicate effectively: Speak the language of your audience, listen carefully to understand the problem you want to solve. Master the art of presenting well: Tell stories that inspire action, get your message across through a combination of data storytelling, negotiation and persuasion skills Be a problem solver: Do your independent research, get your hands dirty and dive deep for answers. Develop your management capabilities: Manage, mentor and help other data scientists reach their full potential.

With the growing demand in ML applications, there is also a demand for machine learning tasks such as data preprocessing, optimizing model hyperparameters and so on to be easily handled by non-experts. This is because, these tasks were repetitive and due to the complexity were considered to be handled only by ML experts. To support this cause and to maintain off-the-shelf quality of machine learning methods without expert knowledge, Google came out with a project named AutoML, an approach that automates designing of ML models. You could also refer to our article on Automated Machine Learning (AutoML) for a clear understanding on how AutoML functions. Trying AutoML on smaller datasets AutoML brought in altogether new dimensions within machine learning workflows where repetitive tasks performed by human experts could be taken over by machines. When Google started off with AutoML, they applied the AutoML approach onto two smaller datasets in DL namely, CIFAR-10 and Penn Treebank to test them on image recognition and language modeling tasks respectively. The result was, AutoML approach could design models that were at par with the ones designed by the ML experts. Also, on comparing the designs drafted by humans and AutoML, it was seen that the machine-suggested architecture included new elements. These elements were later known to alleviate gradient vanishing/exploding issues, which concludes that the machines provided a new architecture which could be more useful for multiple tasks. Also, the machine designed architecture has many channels so that the gradients could flow backwards. This could help explain why LSTM RNNs work better than standard RNNs. Trying AutoML on larger datasets After a success in small scale datasets, Google tested AutoML on large scale datasets such as ImageNet and COCO object detection dataset. Testing AutoML on these was a challenge because of their higher orders of magnitude, and also because simply applying AutoML directly to ImageNet would require many months of training the AutoML method. In order to apply AutoML to large scale datasets, some alterations were made within the AutoML approach for it to be more tractable to large scale datasets. The changes include: Redesigning the search space so that AutoML could find the best layer which can then be stacked many times in a flexible manner to create a final network. Carry out architecture search on CIFAR-10 dataset and transfer the best learned architecture to ImageNet image classification and COCO object detection datasets. Thus, AutoML could find out two best layers i.e normal cell and reduction cell, which when combined resulted into a novel architecture called as “NASNet”. These two work well with CIFAR-10, and also ImageNet and COCO object detection. NASNet was seen to have a prediction accuracy of 82.7% on the validation, as stated by Google. Such an accuracy surpassed all previous inception models built by Google. Further, the learned features from the ImageNet classification were transferred to carry out object detection tasks using the COCO dataset. The learned features combined with a faster R-CNN  resulted into a state-of-the-art predictive performance on the COCO object detection task in both the largest as well as mobile-optimized models. Google suspected that these image features learned by ImageNet and COCO can be reused for various other computer vision applications. Hence, Google open-sourced NASNet for inference on image classification and for object detection in the Slim and Object Detection TensorFlow repositories. Towards Cloud AutoML: Automated Machine learning platform for everyone Cloud AutoML has been Google’s latest buzz for its customers as it makes AI available for everyone. Using Google’s advanced techniques such as learning2learn and transfer learning, Cloud AutoML helps businesses having limited ML expertise, to start building their own high-quality custom models. Thus, Cloud AutoML benefits AI experts by improving their productivity and explore new fields in AI. The experts can also aid less-skilled engineers to build powerful systems. Companies such as Disney and Urban Outfitters are using AutoML for making search and shopping on their websites more relevant. With AutoML going on cloud, Google released its first Cloud AutoML product, Cloud AutoML Vision, an Image Recognition tool that enables fast and easy to build custom ML models. This tool has a drag-and-drop interface that allows one to easily upload images, train and manage the models, and then deploy those trained models directly on Google Cloud. When used to classify popular public datasets like ImageNet and CIFAR, Cloud AutoML Vision  has shown state-of-the-art results. These results included fewer misclassifications than the generic ML APIs results.    Here are some highlights on Cloud AutoML vision: It is built on Google’s leading image recognition approaches, along with transfer learning and neural architecture search technologies. Hence, one can expect an accurate model even if the business has a limited expertise in ML. One can build a simple model in minutes or a full, production-ready model in a day in order to pilot AI-enabled application. AutoML Vision has a simple graphical UI using which one can easily specify data. It later turns the data into a high quality model customized for one’s specific needs. Starting off with Images, Google plans to roll out Cloud AutoML tools and services for text and audio too. However, Google isn’t the only one in the race; other competitors including AWS and Microsoft are also bringing in tools such as Amazon’s SageMaker and Microsoft’s service for customizing Image recognition model, to aid developers with automating machine learning. Some other automated tools include: Auto-sklearn: An automated project that aids scikit-learn project--package of common machine learning functions--to choose the right estimator function. The Auto-sklearn includes a generic estimator function that conducts analysis to determine the best algorithm and set of hyperparameters for a given Scikit-learn job. Auto-WEKA : An inspiration from the Auto-sklearn is for machine learners using Java programming language and the Weka ML package. Auto-WEKA uses a fully automated approach to select a learning algorithm and sets its hyperparameters, unlike previous methods which used to address this in isolation. H2o Driverless AI : This uses a web-based UI and is specifically designed for business users who want to gain insights from data but do not want to get into the intricacies of machine learning algorithms. This tool allows users to choose one or multiple target variables in the dataset that needs a solution, and the system provides the answer. The results are in the form of interactive charts, explained with annotations in plain English. Currently, Google’s AutoML is leading them. It would be exciting to see how Google scales an automated ML environment exactly the same as traditional ML.   Not only Google, but also other businesses are contributing to the movement towards adopting an automated machine learning ecosystem. We saw some tools joining the automation league and can expect more tools to join them. Also, these tools could go on cloud in future for an extended availability for non-experts, similar to the AutoML cloud by Google. With machine learning going automated, we can expect more and more systems to move a step closer to widening the scope for AI.  

In a research published in the Nature journal on Monday, a team of neuroscientists from the University of California, San Francisco, introduced a neural decoder that can synthesize natural-sounding speech based on brain activity. This research was led by Gopala Anumanchipalli, a speech scientist, and Josh Chartier, a bioengineering graduate student in the Chang lab. It is being developed in the laboratory of Edward Chang, a Neurological Surgery professor at University of California. Why is this neural decoder being introduced? There are many cases of people losing their voice because of stroke, traumatic brain injury, or neurodegenerative diseases such as Parkinson’s disease, multiple sclerosis, and amyotrophic lateral sclerosis. Currently,assistive devices that track very small eye or facial muscle movements to enable people with severe speech disabilities express their thoughts by writing them letter-by-letter, do exist. However, generating text or synthesized speech with such devices is often time consuming, laborious, and error-prone. Another limitation these devices have is that they only permit generating a maximum of 10 words per minute, compared to the 100 to 150 words per minute of natural speech. This research shows that it is possible to generate a synthesized version of a person’s voice that can be controlled by their brain activity. The researchers believe that in future, this device could be used to enable individuals with severe speech disability to have fluent communication. It could even reproduce some of the “musicality” of the human voice that expresses the speaker’s emotions and personality. “For the first time, this study demonstrates that we can generate entire spoken sentences based on an individual’s brain activity,” said Chang. “This is an exhilarating proof of principle that with technology that is already within reach, we should be able to build a device that is clinically viable in patients with speech loss.” How does this system work? This research is based on another study by Josh Chartier and Gopala K. Anumanchipalli, which shows how the speech centers in our brain choreograph the movements of the lips, jaw, tongue, and other vocal tract components to produce fluent speech. In this new study, Anumanchipalli and Chartier asked five patients being treated at the UCSF Epilepsy Center to read several sentences aloud. These patients had electrodes implanted into their brains to map the source of their seizures in preparation for neurosurgery. Simultaneously, the researchers recorded activity from a brain region known to be involved in language production. The researchers used the audio recordings of volunteer’s voice to understand the vocal tract movements needed to produce those sounds. With this detailed map of sound to anatomy in hand, the scientists created a realistic virtual vocal tract for each volunteer that could be controlled by their brain activity. The system comprised of two neural networks: A decoder for transforming brain activity patterns produced during speech into movements of the virtual vocal tract. A synthesizer for converting these vocal tract movements into a synthetic approximation of the volunteer’s voice. Here’s a video depicting the working of this system: https://www.youtube.com/watch?v=kbX9FLJ6WKw&feature=youtu.be The researchers observed that the synthetic speech produced by this system was much better as compared to the synthetic speech directly decoded from the volunteer’s brain activity. The generated sentences were also understandable to hundreds of human listeners in crowdsourced transcription tests conducted on the Amazon Mechanical Turk platform. The system is still in its early stages. Explaining its limitations, Chartier said, “We still have a ways to go to perfectly mimic spoken language. We’re quite good at synthesizing slower speech sounds like ‘sh’ and ‘z’ as well as maintaining the rhythms and intonations of speech and the speaker’s gender and identity, but some of the more abrupt sounds like ‘b’s and ‘p’s get a bit fuzzy. Still, the levels of accuracy we produced here would be an amazing improvement in real-time communication compared to what’s currently available.” Read the full report on UCSF’s official website. OpenAI introduces MuseNet: A deep neural network for generating musical compositions Interpretation of Functional APIs in Deep Neural Networks by Rowel Atienza Google open-sources GPipe, a pipeline parallelism Library to scale up Deep Neural Network training  

Continuing the recent trend of rapid updates introducing significant fixes and new features, Google have released the first release candidate for Tensorflow 1.10. TensorFlow 1.10 RC0 brings some improvements in model training and evaluation, and also how Tensorflow runs in a local environment. This is Tensorflow’s fifth update release in just over a month, which includes two major version updates, the previous one being Tensorflow 1.9 What’s new in Tensorflow 1.10 RC0? The tf.contrib.distributions module will be deprecated in this version. This module is primarily used to work with statistical distributions Upgrade to NCCL  2.2 will be mandatory in order to perform GPU computing with this version of Tensorflow, for added performance and efficiency. Model training speed can now be optimized by improving the communication between the model and the Tensorflow resources. For this, the RunConfig function has been updated in this version. The Tensorflow development team also announced support for Bazel - a popular build and testing automation software - and deprecated support for cmake starting with Tensorflow 1.11. This version also incorporated some bug fixes and performance improvements to the tf.data, tf.estimator and other related modules. To get full details on the features list of this release candidate, you can check out Tensorflow’s official release page on Github. No news on Tensorflow 2.0 yet Many developers were expecting the next major release of Tensorflow, Tensorflow 2.0, to be released in late July or August. However, the announcement of this release candidate and the mention of the next version update (1.11) means they will have to wait for some more time before they get to know more about the next breakthrough release. Read more Why Twitter (finally!) migrated to Tensorflow Python, Tensorflow, Excel and more – Data professionals reveal their top tools Can a production ready Pytorch 1.0 give TensorFlow a tough time?

A move that has surprised many, OpenAI yesterday announced the creation of a new for-profit company to balance its huge expenditures into compute and AI talents. Sam Altman, the former president of Y Combinator who stepped down last week, has been named CEO of the new “capped-profit” company, OpenAI LP. But some worry that this move may result in making the innovative company no different from the other AI startups out there. With the OpenAI LP their mission is to ensure that artificial general intelligence (AGI) benefits all of humanity, primarily by attempting to build safe AGI and share the benefits with the world. OpenAI mentions on their blog that “returns for our first round of investors are capped at 100x their investment (commensurate with the risks in front of us), and we expect this multiple to be lower for future rounds as we make further progress.” Any returns beyond the cap amount will revert to OpenAI. OpenAI LP’s primary obligation is to advance the aims of the OpenAI Charter. All investors and employees sign agreements that OpenAI LP’s obligation to the Charter always comes first, even at the expense of some or all of their financial stake. But the major reason behind the new for-profit subsidiary can be explicitly put up as OpenAI in need of more money. The company anticipates to spend billions of dollars in building large-scale cloud compute, attracting and retaining talented people, and developing AI supercomputers in the coming years. The cash burn rate of a top AI research company is staggering. Consider OpenAI’s recent OpenAI Five project — a set of coordinated AI bots trained to compete against human professionals in the video game Dota 2. OpenAI rented 128,000 CPU cores and 256 GPUs at approximately US$2500 per hour for the time-consuming process of training and fine-tuning its OpenAI Five models. Additionally consider the skyrocketing cost of retaining top AI talents. A New York Times story revealed that OpenAI paid its Chief Scientist Ilya Sutskever more than US$1.9 million in 2016. The company currently employs some 100 pricey talents for developing its AI capabilities, safety, and policies. OpenAI LP will be governed by the original OpenAI Board. Only a few on the Board of Directors are allowed to hold financial stakes, and those who do not will be able to vote on decisions if the financial interests are seen to conflict with OpenAI’s mission. People have linked the new for-profit company with OpenAI’s recent controversial decision to withhold the code and training dataset for their language model GPT-2, ostensibly due concerns they might be used for malicious purposes such as generating fake news. A tweet from a software engineer suggested an ulterior motive: “I now see why you didn’t release the fully trained model of #gpt2”. OpenAI Chairman and CTO Greg Brockman shot back: “Nope. We aren’t going to commercialize GPT-2.” OpenAI aims to forge a sustainable path towards long-term AI development. And it also plans to strike a balance between benefiting humanity and turning a profit. A big part of OpenAI’s appeal to top AI talents is it's not-for-profit character — will OpenAI LP mar that? And can OpenAI really strike a balance between benefiting humanity and turning a profit? Whether the for-profit shift will accelerate OpenAI’s mission or prove a detrimental detour remains to be seen, but the journey ahead is bound to be challenging. OpenAI’s new versatile AI model, GPT-2 can efficiently write convincing fake news from just a few words  

Last week, Google researchers announced the release of  Google Research Football Environment, a reinforcement learning environment where agents can master football. This environment comes with a physics-based 3D football simulation where agents control either one or all football players on their team, they learn how to pass between them, and further manage to overcome their opponent’s defense to score goals. The Football Environment offers a game engine, a set of research problems called Football Benchmarks and Football Academy and much more. The researchers have released a beta version of open-source code on Github to facilitate the research. Let’s have a brief look at each of the elements in the Google Research Football Environment. Football engine: The core of the Football Environment Based on the modified version of Gameplay Football, the Football engine simulates a football match including fouls, goals, corner and penalty kicks, and offsides. The engine is programmed in C++,  which allows it to run with GPU as well as without GPU-based rendering enabled. The engine allows learning from different state representations that contain semantic information such as the player’s locations and learning from raw pixels. The engine can be run in both stochastic mode as well as deterministic mode for investigating the impact of randomness. The engine is also compatible with OpenAI Gym API. Read Also: Create your first OpenAI Gym environment [Tutorial] Football Benchmarks: Learning from the actual field game The researchers propose a set of benchmark problems for RL research based on the Football Engine with the help of Football Benchmarks. These benchmarks highlight the goals such as playing a “standard” game of football against a fixed rule-based opponent. The researchers have provided three versions, the Football Easy Benchmark, the Football Medium Benchmark, and the Football Hard Benchmark, which differ only in the strength of the opponent. They also provide benchmark results for two state-of-the-art reinforcement learning algorithms including DQN and IMPALA that can be run in multiple processes on a single machine or concurrently on many machines. Image Source: Google’s blog post These results indicate that the Football Benchmarks are research problems that vary in difficulties. According to the researchers, the Football Easy Benchmark is suitable for research on single-machine algorithms and Football Hard Benchmark is challenging for massively distributed RL algorithms. Football Academy: Learning from a set of difficult scenarios   Football Academy is a diverse set of scenarios of varying difficulty that allow researchers to look into new research ideas and allow testing of high-level concepts. It also provides a foundation for investigating curriculum learning, research ideas, where agents can learn harder scenarios. The official blog post states, “Examples of the Football Academy scenarios include settings where agents have to learn how to score against the empty goal, where they have to learn how to quickly pass between players, and where they have to learn how to execute a counter-attack. Using a simple API, researchers can further define their own scenarios and train agents to solve them.” Users are giving mixed reaction to this news as some find nothing new in Google Research Football Environment. A user commented on HackerNews, “I guess I don't get it... What does this game have that SC2/Dota doesn't? As far as I can tell, the main goal for reinforcement learning is to make it so that it doesn't take 10k learning sessions to learn what a human can learn in a single session, and to make self-training without guiding scenarios feasible.” Another user commented, “This doesn't seem that impressive: much more complex games run at that frame rate? FIFA games from the 90s don't look much worse and certainly achieved those frame rates on much older hardware.” While a few others think that they can learn a lot from this environment. Another comment reads, “In other words, you can perform different kinds of experiments and learn different things by studying this environment.” Here’s a short YouTube video demonstrating Google Research Football. https://youtu.be/F8DcgFDT9sc To know more about this news, check out Google’s blog post. Google researchers propose building service robots with reinforcement learning to help people with mobility impairment Researchers propose a reinforcement learning method that can hack Google reCAPTCHA v3 Researchers input rabbit-duck illusion to Google Cloud Vision API and conclude it shows orientation-bias  

On Monday, Intel Labs introduced first of its kind cryogenic control chip codenamed Horse Ridge. According to Intel, Horse Ridge will enable commercially viable quantum computers and speed up development of full-stack quantum computing systems. Intel announced that Horse Ridge will enable control of multiple quantum bits (qubits) and set a clear path toward scaling larger systems. This seems to be a major milestone on the path to quantum practicality. As right now the challenge for quantum computing is that it only works at near-freezing temperatures. Intel is trying to change that with this control chip. As per Intel, Horse Ridge will be able to enable control at very low temperatures, as it will eliminate hundreds of wires going into a refrigerated case that houses the quantum computer. Horse Ridge is developed in partnership with Intel’s research collaborators at QuTech at Delft University of Technology. It is fabricated using Intel’s 22-nanometer FinFET manufacturing technology. The in-house fabrication of these control chips at Intel will dramatically accelerate the company’s ability to design, test, and optimize a commercially viable quantum computer, the company said. “A lot of research has gone into qubits, which can do simultaneous calculations. But Intel saw that controlling the qubits created another big challenge to developing large-scale commercial quantum systems,” states Jim Clarke, Director of quantum hardware, Intel in the official press release . “It’s pretty unique in the community, as we’re going to take all these racks of electronics you see in a university lab and miniaturize that with our 22-nanometer technology and put it inside of a fridge,” added Clarke. “And so we’re starting to control our qubits very locally without having a lot of complex wires for cooling.” The name “Horse Ridge” is inspired from one of the coldest regions in Oregon known as the Horse Ridge. It is designed to operate at cryogenic temperatures, approx 4 degrees Kelvin which is 7 degrees Fahrenheit and 4 degrees Celsius. What is the innovation behind Horse Ridge Quantum computers promise the potential to tackle problems that conventional computers can’t handle by themselves. Quantum computers leverage a phenomenon of quantum physics that allows qubits to exist in multiple states simultaneously. As a result, qubits can conduct a large number of calculations at the same time dramatically speeding up complex problem-solving. But Intel acknowledges the fact that the quantum research community still lags behind in demonstrating quantum practicality, a benchmark to determine if a quantum system can deliver game-changing performance to solve real-world problems. Till date, researchers have focused on building small-scale quantum systems to demonstrate the potential of quantum devices. In these efforts, researchers have relied upon existing electronic tools and high-performance computing rack-scale instruments to connect the quantum system to the traditional computational devices that regulates qubit performance and programs the system inside the cryogenic refrigerator. These devices are often custom designed to control individual qubits, requiring hundreds of connective wires in and out of the refrigerator. However, this extensive control cabling for each qubit hinders the ability to scale the quantum system to the hundreds or thousands of qubits required to demonstrate quantum practicality, not to mention the millions of qubits required for a commercially viable quantum solution. With Horse Ridge, Intel radically simplifies the control electronics required to operate a quantum system. Replacing these bulky instruments with a highly integrated system-on-chip (SoC) will simplify system design and allow for sophisticated signal processing techniques to accelerate set-up time, improve qubit performance, and enable the system to efficiently scale to larger qubit counts. “One option is to run the control electronics at room temperature and run coax cables down to configure the qubits. But you can immediately see that you’re going to run into a scaling problem because you get to hundreds or thousands of cables and it’s not going to work,” said Richard Uhlig, Managing Director Intel Labs. “What we’ve done with Horse Ridge is that it’s able to run at temperatures that are much closer to the qubits themselves. It runs at about 4 degrees Kelvin. The innovation is that we solved the challenges around getting CMOS to run at those temperatures and still have a lot of flexibility in how the qubits are controlled and configured.” To know more about this exciting news, check out the official announcement from Intel. Are we entering the quantum computing era? Google’s Sycamore achieves ‘quantum supremacy’ while IBM refutes the claim The US to invest over $1B in quantum computing, President Trump signs a law Quantum computing, edge analytics, and meta learning: key trends in data science and big data in 2019

After a long wait, Intel has officially launched its first 10th generation core processors, code-named ‘Ice Lake’. The first batch contains 11 highly integrated 10nm processors which showcases high-performance artificial intelligence (AI) features and is designed for sleek 2 in 1s and laptops. The ‘Ice Lake’ processors are manufactured on Intel’s 10nm processor and consist of the 14nm chipset in the same carrier. It includes two or four Sunny Cove cores along with Intel’s Gen 11 Graphics processing unit (GPU). The 10nm measure of the processor indicates the size of the transistors used. The 10 nanometer miniscule length also shows the power of the transistor as it is considered that smaller the transistor, better is its power consumption. Read More: Intel unveils the first 3D Logic Chip packaging technology, ‘Foveros’, powering its new 10nm chips, ‘Sunny Cove’ Chris Walker, Intel corporate vice president and general manager of Mobility Client Platforms in the Client Computing Group says that “With broad-scale AI for the first time on PCs, an all-new graphics architecture, best-in-class Wi-Fi 6 (Gig+) and Thunderbolt 3 – all integrated onto the SoC, thanks to Intel’s 10nm process technology and architecture design – we’re opening the door to an entirely new range of experiences and innovations for the laptop.” Intel was supposed to ship the 10nm processors, way back in 2016. Intel CEO Bob Swan says that the delay was due to the “company’s overly aggressive strategy for moving to its next node.” Intel has also introduced a new processor number naming structure for the 10th generation ‘Ice Lake’ processors which indicates the generation and the level of graphics performance of the processor. Image source: Intel What’s new in the 10th generation Intel core processors? Intelligent performance The 10th generation core processors are the first purpose-built processors for AI on laptops and 2 in 1s. They are built for modern AI-infused applications and contains many features such as: Intel Deep Learning Boost, used for specifically boosting flexibility to run complex AI workloads. It has a dedicated instruction set that accelerates neural networks on the CPU for maximum responsiveness. Up to 1 teraflop of GPU engine compute for sustained high-throughput inference applications Intel’s Gaussian & Neural Accelerator (GNA) provides an exclusive engine for background workloads such as voice processing and noise prevention at ultra-low power, for utmost battery life. New graphics With the Iris Plus graphics, the 10th generation core processors imparts double graphic performance in 1080p and higher-level content creation in 4K video editing, application of video filters and high-resolution photo processing. This is the first time that Intel’s Graphics processing unit (GPU) will support VESA’s Adaptive Sync* display standard. It enables a smoother gaming experience across games like Dirt Rally 2.0* and Fortnite*. According to Intel, this is the industry's first integrated GPU to incorporate variable rate shading for better rendering performance, as it uses the Gen11 graphics architecture.  The 10th generation core processors supports the BT.2020* specification, hence it is possible to view a 4K HDR video in a billion colors. Best connectivity With improved board integration, PC manufacturers can innovate on form factor for sleeker designs with Wi-Fi 6 (Gig+) connectivity and up to four Thunderbolt 3 ports. Intel claims this is the “fastest and most versatile USB-C connector available.” In the first batch of 11 'Ice Lake' processors, there are 6 Ice Lake U series and 5 Ice Lake Y series processors. Given below is the complete Ice Lake processors list. Image Source: Intel Intel has revealed that laptops with the 10th generation core processors can be expected in the holiday season this year. The post also states that they will soon release additional products in the 10th generation Intel core mobile processor family due to increased needs in computing. The upcoming processors will “deliver increased productivity and performance scaling for demanding, multithreaded workloads.”   Users love the new 10th generation core processor features and are especially excited about the Gen 11 graphics. https://twitter.com/Tribesigns/status/1133284822548279296 https://twitter.com/Isaacraft123/status/1156982456408596481 Many users are also expecting to see the new processors in the upcoming Mac notebooks. https://twitter.com/ChernSchwinn1/status/1157297037336928256 https://twitter.com/matthewmspace/status/1157295582844575744 Head over to the Intel newsroom page for more details. Apple advanced talks with Intel to buy its smartphone modem chip business for $1 billion, reports WSJ Why Intel is betting on BFLOAT16 to be a game changer for deep learning training? Hint: Range trumps Precision. Intel’s new brain inspired neuromorphic AI chip contains 8 million neurons, processes data 1K times faster

Climate change effects are now visible in all countries around the globe. The world is witnessing phenomena like higher temperature, flooding, ice-melting, and much more. There have been many technologies invented in the last decade to help humans understand and adapt to these climatic changes. Earlier this month, researchers from Montreal Institute for Learning Algorithms, ConscientAI Labs and Microsoft Research came up with a project that aims to generate images which will depict accurate, vivid, and personalized outcomes of climate change using Machine Learning (ML) and Cycle-Consistent Adversarial Networks (CycleGANs). This will enable individuals to make more informed choices about their climate future by creating an understanding of the effects of climate change, while maintaining scientific credibility using climate model projections. The project is to develop a Machine Learning (ML) based tool. This tool will show in a personalized way, the probable effect that climate change will have on a specific location familiar to the viewer. When given an address, the tool will generate an image projecting transformations which are likely to occur there, based on a formal climate model. For the initial version, the generated images consist of houses and buildings specifically after flooding events. The challenge in generating realistic images using Cycle-Consistent Adversarial Networks (CycleGANs) is to collect the training data needed in order to extract the mapping function. The researchers manually searched open source photo-sharing websites for images of houses from various neighborhoods and settings, such as suburban detached houses, urban townhouses, and apartment buildings. They have gathered over 500 images of non-flooded houses and the same number of flooded locations, and re-sized them to 300x300 pixels. The networks were trained using the publicly available PyTorch. The CycleGAN model for 200 epochs were trained using the trained images, using the Adam solver with a batch size of 1 and training the model from scratch with a learning rate of 0.0002. As per the CycleGAN training procedure, the learning rate is constant for the first 100 epochs and linearly decayed to zero over the next 100 epochs. Project Output and Future Plan The trained CycleGAN model was successful in learning an adequate mapping between grass and water, which could be applied to generate fairly realistic images of flooded houses. This will work best with single-family, suburban-type houses which are surrounded by an expanse of grass. From the 80 images in the test set, it was found that about 70% were successfully mapped to realistically flooded houses. This initial version of the CycleGAN model will illustrate the feasibility of applying generative model to create personalized images of an extreme climate event i.e., flooding, that is expected to increase in frequency based on climate change projections. Subsequent versions of this model will integrate more varied types of houses and surroundings, as well as different types of climate-change related extreme event phenomena (i.e. droughts, hurricanes, wildfires, air pollution etc), depending on the expected impacts at a given location, as well as forecast time horizons. There’s still scope for improvement with regard to the color scheme of the generated images and the visual artifacts. Furthermore to channel the emotional response of the public, into behavioural change or actions, the researchers are planning another improvement to the model called ‘choice knobs’. This will enable users to visually see the impact of their personal choices, such as deciding to use more public transportation, as well as the impact of broader policy decisions, such as carbon tax and increasing renewable portfolio standards. The projects greater aim is to help the general population progress towards more visible public support for climate change mitigation steps on a national level, facilitating governmental interventions and helping make the required rapid changes to a global sustainable economy. The researchers have stated that they need to explore more physical constraints to GAN training in order to incorporate more physical knowledge into these projections. This will enable a GAN model to transform a house to its projected flooded state and also take into account the forecast simulations of the flooding event represented by the physical variable outputs and probabilistic scenarios by a climate model for a given location. Response to the project Few developers have liked the idea of using technology, to produce realistic images depicting the effect of climate change in your own hometown which may make people understand the adverse effects of it. https://twitter.com/jameskobielus/status/1129392932988096513 While some developers are not sure if showing people a picture of their house submerged in water is going to create any difference. A user on Hacker news comments, “The threshold for believing the effects of climate change has to change from reading/seeing to actually being there and touching it. Or some far more reliable system of remote verification has to be established” Another user adds, “Is this a real paper? It's got to be a joke, right? a parody? It's literally a request to develop images to be used for propaganda purposes. And for those who will say that climate change is going to end the world, yeah, but that doesn't mean we should develop propaganda technology that could be used for some other political purpose.” There are already many studies/evidences to make people aware of the effects of climate change, depicting a picture of their house submerged in water is not going to move them anymore. Climate change is already happening and effecting our day to day lives. What we need now are stronger approaches towards analysing, mitigating, and adapting to these changes and inspiring more government policies to fight against these climate changes. To know more details about the project, head over to the research paper. Read More Amazon employees get support from Glass Lewis and ISS on its resolution for Amazon to manage climate change risks ICLR 2019 Highlights: Algorithmic fairness, AI for social good, climate change, protein structures, GAN magic, adversarial ML and much more Responsible tech leadership or climate washing? Microsoft hikes its carbon tax and announces new initiatives to tackle climate change

Last week, researchers from the Imperial College in London and Samsung’s AI research center in the UK revealed how deepfakes can be used to generate a singing or talking video portrait by from a still image of a person and an audio clip containing speech. In their paper titled, “Realistic Speech-Driven Facial Animation with GANs”, the researchers have used temporal GAN which uses 3 discriminators focused on achieving detailed frames, audio-visual synchronization, and realistic expressions. Source: arxiv.org “The generated videos are evaluated based on sharpness, reconstruction quality, lip-reading accuracy, synchronization as well as their ability to generate natural blinks”, the researchers mention in their paper. https://youtu.be/9Ctm4rTdVTU Researchers used the GRID, TCD TIMIT, CREMA-D and LRW datasets. The GRID dataset has 33 speakers each uttering 1000 short phrases, containing 6 words randomly chosen from a limited dictionary. The TCD TIMIT dataset has 59 speakers uttering approximately 100 phonetically rich sentences each. The CREMA-D dataset includes 91 actors coming from a variety of different age groups and races utter 12 sentences. Each sentence is acted out by the actors multiple times for different emotions and intensities. Researchers have used the recommended data split for the TCD TIMIT dataset but exclude some of the test speakers and use them as a validation set. Researchers performed data augmentation on the training set by mirroring the videos. Metrics used to assess the quality of generated videos Researchers evaluated the videos using traditional image reconstruction and sharpness metrics. These metrics can be used to determine frame quality; however, they fail to reflect other important aspects of the video such as audio-visual synchrony and the realism of facial expressions. Hence they have also proposed alternative methods capable of capturing these aspects of the generated videos. Reconstruction Metrics This method uses common reconstruction metrics such as the peak signal-to-noise ratio (PSNR) and the structural similarity (SSIM) index to evaluate the generated videos. However, the researchers reveal that “reconstruction metrics will penalize videos for any facial expression that does not match those in the ground truth videos”. Sharpness Metrics The frame sharpness is evaluated using the cumulative probability blur detection (CPBD) measure, which determines blur based on the presence of edges in the image. For this metric as well as for the reconstruction metrics larger values imply better quality. Content Metrics The content of the videos is evaluated based on how well the video captures the identity of the target and on the accuracy of the spoken words. The researchers have verified the identity of the speaker using the average content distance (ACD), which measures the average Euclidean distance of the still image representation, obtained using OpenFace from the representation of the generated frames. The accuracy of the spoken message is measured using the word error rate (WER) achieved by a pre-trained lip-reading model. They used the LipNet model which exceeds the performance of human lip-readers on the GRID dataset. For both content metrics, lower values indicate better accuracy. Audio-Visual Synchrony Metrics Synchrony is quantified in Joon Son Chung and Andrew Zisserman’s “Out of time: automated lip sync in the wild”. In this work Chung et al. propose the SyncNet network which calculates the euclidean distance between the audio and video encodings on small (0.2 second) sections of the video. The audio-visual offset is obtained by using a sliding window approach to find where the distance is minimized. The offset is measured in frames and is positive when the audio leads the video. For audio and video pairs that correspond to the same content, the distance will increase on either side of the point where the minimum distance occurs. However, for uncorrelated audio and video, the distance is expected to be stable. Based on this fluctuation they further propose using the difference between the minimum and the median of the Euclidean distances as an audio-visual (AV) confidence score which determines the audio-visual correlation. Higher scores indicate a stronger correlation, whereas confidence scores smaller than 0.5 indicate that Limitations and the possible misuse of Deepfake The limitation of this new Deepfake method is that it only works for well-aligned frontal faces. “the natural progression of this work will be to produce videos that simulate in wild conditions”, the researchers mention. While this research appears the next milestone for GANs in generating videos from still photos, it also may be misused for spreading misinformation by morphing video content from any still photograph. Recently, at the House Intelligence Committee hearing, Top House Democrat Rep. Adam Schiff (D-CA) issued a warning on Thursday that deepfake videos could have a disastrous effect on the 2020 election cycle. “Now is the time for social media companies to put in place policies to protect users from this kind of misinformation not in 2021 after viral deepfakes have polluted the 2020 elections,” Schiff said. “By then it will be too late.” The hearing came only a few weeks after a real-life instance of a doctored political video, where the footage was edited to make House Speaker Nancy Pelosi appear drunk, that spread widely on social media. “Every platform responded to the video differently, with YouTube removing the content, Facebook leaving it up while directing users to coverage debunking it, and Twitter simply letting it stand,” The Verge reports. YouTube took the video down; however, Facebook refused to remove the video. Neil Potts, Public Policy Director of Facebook had stated that if someone posted a doctored video of Zuckerberg, like one of Pelosi, it would stay up. After this, on June 11, a fake video of Mark Zuckerberg was posted on Instagram, under the username, bill_posters_uk. In the video, Zuckerberg appears to give a threatening speech about the power of Facebook. https://twitter.com/motherboard/status/1138536366969688064 Omer Ben-Ami, one of the founders of Canny says that the video is made to educate the public on the uses of AI and to make them realize the potential of AI. Though Zuckerberg’s video was to retain the educational value of Deepfakes, this shows the potential of how it can be misused. Although some users say it has interesting applications, many are concerned that the chances of misusing this software are more than putting it into the right use. https://twitter.com/timkmak/status/1141784420090863616 A user commented on Reddit, “It has some really cool applications though. For example in your favorite voice acted video game, if all of the characters lips would be in sync with the vocals no matter what language you are playing the game in, without spending tons of money having animators animate the characters for every vocalization.” To know more about this new Deepfake, read the official research paper. Lawmakers introduce new Consumer privacy bill and Malicious Deep Fake Prohibition Act to support consumer privacy and battle deepfakes Worried about Deepfakes? Check out the new algorithm that manipulate talking-head videos by altering the transcripts Machine generated videos like Deepfakes – Trick or Treat?

The Python team came out with a new simple and powerful library called Pypeline, last week for creating concurrent data pipelines. Pypeline has been designed for solving simple to medium data tasks that require concurrency and parallelism. It can be used in places where using frameworks such as Spark or Dask feel unnatural. Pypeline comprises an easy to use familiar and functional API. It enables building data pipelines using Processes, Threads, and asyncio.Tasks via the exact same API. With Pypeline, you also have control over memory and CPU resources which are used at each stage of your pipeline. Pypeline Basic Usage Using Pypeline, you can easily create multi-stage data pipelines with the help of functions such as map, flat_map, filter, etc. To do so, you need to define a computational graph specifying the operations which are to be performed at each stage, the number of resources, and the type of workers you want to use. Pypeline comes with 3 main modules, and each of them uses a different type of worker. To build multi-stage data pipelines, you can use 3 type of workers, namely, processes, threads, and tasks. Processes You can create a pipeline based on multiprocessing. Process workers with the help of process module. After this, you can specify the numbers of workers at each stage. The maxsize parameter limits the maximum amount of elements that the stage can hold simultaneously. Threads and Tasks Create a pipeline using threading.Thread workers by using the thread module. Additionally, in order to create a pipeline based on asyncio.Task workers, use an asyncio_task module. Apart from being used to create multi-stage data pipelines, it can also help you create pipelines with the help of the pipe | operator. For more information, check out the official documentation. How to build a real-time data pipeline for web developers – Part 1 [Tutorial] How to build a real-time data pipeline for web developers – Part 2 [Tutorial] Create machine learning pipelines using unsupervised AutoML [Tutorial]

Almost after a week of Microsoft’s announcement about its plan to develop a computer vision develop kit for edge computing, Intel smartly introduced its latest offering, called OpenVINO in the domain of Internet of Things (IoT) and Artificial Intelligence (AI). This toolkit is a comprehensive computer vision solution, that brings computer vision and deep learning capabilities to the edge devices smoothly. OpenVINO (Open Visual Inference and Neural Network Optimization) toolkit supports popular open source frameworks like OpenCV, Caffe and TensorFlow. It supports and works with Intel’s traditional CPUs, AI chips, field programmable gate array (FPGA) chips and Movidius vision processing unit (VPU). The toolkit presumes the potential to address a wide number of challenges faced by developers in delivering distributed and end-to-end intelligence. With OpenVINO, developers can simply streamline their deep learning inferences and deploy high-performance computer vision solutions across a wide range of use-cases. Computer vision limitations related to bandwidth, latency and storage are expected to be resolved to an extent. This toolkit would also help developers in optimizing AI-integrated computer vision applications and scaling distributed vision applications which generally needs a complete redesign of solution. Until now, edge computing has been more of a prospect for an IoT market. With OpenVINO, Intel stands as the the only industry leader in delivering IoT solutions from the edges, providing an unparalleled solution to meet AI needs of businesses. OpenVINO is already being used by companies like GE Healthcare, Dahua, Amazon Web Services and Honeywell across their Digital Imaging and IoT Solutions. To explore more information on its capabilities and performance, visit Intel’s official OpenVINO product documentation. A gentle note to readers: OpenVINO  is not to be confused with Openvino, an open-source winery and wine-backed cryptoasset, Openvino. Should you go with Arduino Uno or Raspberry Pi 3 for your next IoT project? AWS Greengrass brings machine learning to the edge Cognitive IoT: How Artificial Intelligence is remoulding Industrial and Consumer IoT

Deep learning and 3D vision research have led to major developments in the field of robotics and computer graphics. However, there is a dearth of systems that allow easy loading of popular 3D datasets and get the 3D data across various representations converted into modern machine learning frameworks. To overcome this barrier, researchers at NVIDIA have developed a 3D deep learning library for PyTorch called ‘Kaolin’. Last week, the researchers published the details of Kaolin in a paper titled “Kaolin: A PyTorch Library for Accelerating 3D Deep Learning Research”. https://twitter.com/NvidiaAI/status/1194680942536736768 Kaolin provides an efficient implementation of all core modules that are required to build 3D deep learning applications. According to NVIDIA, Kaolin can slash the job of preparing a 3D model for deep learning from 300 lines of code down to just five. Key features offered by Kaolin It supports all popular 3D representations like Polygon meshes, Pointclouds, Voxel grid, Signed distance functions, and Depth images. It enables complex 3D datasets to be loaded into machine-learning frameworks, irrespective of how they’re represented or will be rendered. It can be implemented in diverse fields for instance robotics, self-driving cars, medical imaging, and virtual reality. Kaolin has a suite of 3D geometric functions that allow manipulation of 3D content. Several rigid body transformations can be implemented in a variety of parameterizations like Euler angles, Lie groups, and Quaternions. It also permits differentiable image warping layers and also allows for 3D-2D projection, and 2D-3D back projection. Kaolin reduces the large overhead involved in file handling, parsing, and augmentation into a single function call and renders support to many 3D datasets like ShapeNet and PartNet. The access to all data is provided via extensions to the PyTorch Dataset and DataLoader classes which makes pre-processing and loading 3D data simple and intuitive. Kaolin’s modular differentiable renderer A differentiable renderer is a process that supplies pixels as a function of model parameters to simulate a physical imaging system. It also supplies derivatives of the pixel values with respect to those parameters. With an aim to allow users the easy use of popular differentiable rendering methods, Kaolin provides a flexible and modular differentiable renderer. It defines an abstract base class called ‘DifferentiableRenderer’ which contains abstract methods for each component in a rendering pipeline. The abstract methods allowed in Kaolin include geometric transformations, lighting, shading, rasterization, and projection. It also supports multiple lighting, shading, projection, and rasterization modes. One of the important aspects of any computer vision task is visualizing data. Kaolin delivers visualization support for all of computer vision representation types. It is implemented via lightweight visualization libraries such as Trimesh, and pptk for running time visualization. The researchers say, “While we view Kaolin as a major step in accelerating 3D DL research, the efforts do not stop here. We intend to foster a strong open-source community around Kaolin, and welcome contributions from other 3D deep learning researchers and practitioners.” The researchers are hopeful that the 3D community will try out Kaolin, and contribute to its development. Many developers have expressed interest in the Kaolin PyTorch Library. https://twitter.com/RanaHanocka/status/1194763643700858880 https://twitter.com/AndrewMendez19/status/1194719320951197697 Read the research paper for more details about Kaolin’s roadmap. You can also check out NVIDIA’s official announcement. Facebook releases PyTorch 1.3 with named tensors, PyTorch Mobile, 8-bit model quantization, and more Transformers 2.0: NLP library with deep interoperability between TensorFlow 2.0 and PyTorch, and 32+ pretrained models in 100+ languages Introducing ESPRESSO, an open-source, PyTorch based, end-to-end neural automatic speech recognition (ASR) toolkit for distributed training across GPUs Baidu adds Paddle Lite 2.0, new development kits, EasyDL Pro, and other upgrades to its PaddlePaddle deep learning platform CNCF announces Helm 3, a Kubernetes package manager and tool to manage charts and libraries