A Deep Learning Approach to Data Compression


We introduce Bit-Swap, a scalable and effective lossless data compression technique based on deep learning. It extends previous work on practical compression with latent variable models, based on bits-back coding and asymmetric numeral systems. In our experiments Bit-Swap is able to beat benchmark compressors on a highly diverse collection of images. We’re releasing code for the method and optimized models such that people can explore and advance this line of modern compression ideas. We also release a demo and a pre-trained model for Bit-Swap image compression and decompression on your own image. See the end of the post for a talk that covers how bits-back coding and Bit-Swap works.


Evaluating and Testing Unintended Memorization in Neural Networks


It is important whenever designing new technologies to ask “how will this affect people’s privacy?” This topic is especially important with regard to machine learning, where machine learning models are often trained on sensitive user data and then released to the public. For example, in the last few years we have seen models trained on users’ private emails, text messages, and medical records.

This article covers two aspects of our upcoming USENIX Security paper that investigates to what extent neural networks memorize rare and unique aspects of their training data.

Specifically, we quantitatively study to what extent following problem actually occurs in practice:


Learning to Learn with Probabilistic Task Embeddings


To operate successfully in a complex and changing environment, learning agents must be able to acquire new skills quickly. Humans display remarkable skill in this area — we can learn to recognize a new object from one example, adapt to driving a different car in a matter of minutes, and add a new slang word to our vocabulary after hearing it once. Meta-learning is a promising approach for enabling such capabilities in machines. In this paradigm, the agent adapts to a new task from limited data by leveraging a wealth of experience collected in performing related tasks. For agents that must take actions and collect their own experience, meta-reinforcement learning (meta-RL) holds the promise of enabling fast adaptation to new scenarios. Unfortunately, while the trained policy can adapt quickly to new tasks, the meta-training process requires large amounts of data from a range of training tasks, exacerbating the sample inefficiency that plagues RL algorithms. As a result, existing meta-RL algorithms are largely feasible only in simulated environments. In this post, we’ll briefly survey the current landscape of meta-RL and then introduce a new algorithm called PEARL that drastically improves sample efficiency by orders of magnitude. (Check out the research paper and the code.)


1000x Faster Data Augmentation


Effect of Population Based Augmentation applied to images, which differs at different percentages into training.

In this blog post we introduce Population Based Augmentation (PBA), an algorithm that quickly and efficiently learns a state-of-the-art approach to augmenting data for neural network training. PBA matches the previous best result on CIFAR and SVHN but uses one thousand times less compute, enabling researchers and practitioners to effectively learn new augmentation policies using a single workstation GPU. You can use PBA broadly to improve deep learning performance on image recognition tasks.

We discuss the PBA results from our recent paper and then show how to easily run PBA for yourself on a new data set in the Tune framework.


Autonomous Vehicles for Social Good: Learning to Solve Congestion


We are in the midst of an unprecedented convergence of two rapidly growing trends on our roadways: sharply increasing congestion and the deployment of autonomous vehicles. Year after year, highways get slower and slower: famously, China’s roadways were paralyzed by a two-week long traffic jam in 2010. At the same time as congestion worsens, hundreds of thousands of semi-autonomous vehicles (AVs), which are vehicles with automated distance and lane-keeping capabilities, are being deployed on highways worldwide. The second trend offers a perfect opportunity to alleviate the first. The current generation of AVs, while very far from full autonomy, already hold a multitude of advantages over human drivers that make them perfectly poised to tackle this congestion. Humans are imperfect drivers: accelerating when we shouldn’t, braking aggressively, and make short-sighted decisions, all of which creates and amplifies patterns of congestion.


End-to-End Deep Reinforcement Learning
without Reward Engineering


Communicating the goal of a task to another person is easy: we can use language, show them an image of the desired outcome, point them to a how-to video, or use some combination of all of these. On the other hand, specifying a task to a robot for reinforcement learning requires substantial effort. Most prior work that has applied deep reinforcement learning to real robots makes uses of specialized sensors to obtain rewards or studies tasks where the robot’s internal sensors can be used to measure reward. For example, using thermal cameras for tracking fluids, or purpose-built computer vision systems for tracking objects. Since such instrumentation needs to be done for any new task that we may wish to learn, it poses a significant bottleneck to widespread adoption of reinforcement learning for robotics, and precludes the use of these methods directly in open-world environments that lack this instrumentation.

We have developed an end-to-end method that allows robots to learn from a modest number of images that depict successful completion of a task, without any manual reward engineering. The robot initiates learning from this information alone (around 80 images), and occasionally queries a user for additional labels. In these queries, the robot shows the user an image and asks for a label to determine whether that image represents successful completion of the task or not. We require a small number of such queries (around 25-75), and using these queries, the robot is able to learn directly in the real world in 1-4 hours of interaction time, resulting in one of the most efficient real-world image-based robotic RL methods. We have open-sourced our implementation.

Our method allows us to solve a host of real world robotics problems from pixels in an end-to-end fashion without any hand-engineered reward functions.


Model-Based Reinforcement Learning from Pixels with Structured Latent Variable Models


Imagine a robot trying to learn how to stack blocks and push objects using visual inputs from a camera feed. In order to minimize cost and safety concerns, we want our robot to learn these skills with minimal interaction time, but efficient learning from complex sensory inputs such as images is difficult. This work introduces SOLAR, a new model-based reinforcement learning (RL) method that can learn skills – including manipulation tasks on a real Sawyer robot arm – directly from visual inputs with under an hour of interaction. To our knowledge, SOLAR is the most efficient RL method for solving real world image-based robotics tasks.

Our robot learns to stack a Lego block and push a mug onto a coaster with only inputs from a camera pointed at the robot. Each task takes an hour or less of interaction to learn.


Large-Scale Long-Tailed Recognition in an Open World


Existing Computer Vision Setting v.s. Real-World Scenario

One day, an ecologist came to us. He wanted to use modern computer vision techniques to perform automatic animal identification in his wildlife camera trap image datasets. We were so confident because it sounded just like a basic image classification problem. However, we failed. The dataset he provided was extremely long-tailed and open-ended. As usual, when we did not have enough training data, we asked if it was possible to provide more data for the tail classes and just ignore the open classes that might appear in the testing dataset. Unfortunately, collecting more data was not the option. It could take an extremely long time for these ecologists to take photos of rare and secluded animals in the wild. For some endangered animals, they even had to wait for years for one single shot. At the same time, new animal species kept coming in, and old animal species kept leaving. The total class number was never fixed in such a dynamic system. Moreover, the identification of rare and new animals has more conservational values than abundant animals. If we could only do well on the abundant classes, the method would never be practically usable. We tried all possible methods we could think of (data augmentation, sampling techniques, few-shot learning, imbalanced classification, etc.); but none of the existing methods could handle abundant classes, scarce classes and open classes at the same time (Fig. 1).

Figure 1: There exists a considerable gap between the existing computer vision setting and the real-world scenario.


Robots that Learn to Adapt


Figure 1: Our model-based meta reinforcement learning algorithm enables a legged robot to adapt online in the face of an unexpected system malfunction (note the broken front right leg).

Humans have the ability to seamlessly adapt to changes in their environments: adults can learn to walk on crutches in just a few seconds, people can adapt almost instantaneously to picking up an object that is unexpectedly heavy, and children who can walk on flat ground can quickly adapt their gait to walk uphill without having to relearn how to walk. This adaptation is critical for functioning in the real world.


Robots that Learn to Use Improvised Tools


In many animals, tool-use skills emerge from a combination of observational learning and experimentation. For example, by watching one another, chimpanzees can learn how to use twigs to “fish” for insects. Similarly, capuchin monkeys demonstrate the ability to wield sticks as sweeping tools to pull food closer to themselves. While one might wonder whether these are just illustrations of “monkey see, monkey do,” we believe these tool-use abilities indicate a greater level of intelligence.

Left: A chimpanzee fishing for termites. Right: A gorilla using a stick to gather herbs. (source)

The question our new work explores is: can we enable robots to use tools in the same way — through observation and experimentation?

A requisite for performing complex multi-object manipulation tasks, such as those involved in tool use, is an understanding of physical cause-and-effect relationships. Therefore, the ability to predict how one object might interact with another is crucial. Our prior work has investigated how visual predictive models of cause-and-effect can be learned from unsupervised robot interaction with the world. After learning such a model, the robot can plan to accomplish a diverse set of simple tasks, including cloth folding and object arrangement. However, if we consider the more complex interactions that occur in tool-use tasks, such as how a broom can sweep dirt into a dustpan, undirected experimentation isn’t enough.

Hence, taking inspiration from how animals learn, we designed an algorithm that allows robots to learn tool-use skills through a similar paradigm of imitation and interaction. In particular, we show that, with a mix of demonstration data and unsupervised experience, a robot can use novel objects as tools and even improvise tools in the absence of traditional ones. Further, depending on the demands of the task, our method demonstrates the ability to decide whether to use the provided tools. In this post, we will describe how this works.