Lila Ibrahim is the first ever chief operating officer of DeepMind, one of the world's best known artificial intelligence companies.

I explain Artificial Intelligence terms and news to non-experts. Wildfires are more and more present in modern society, mainly caused by heat waves, lightning, droughts, climate change, or even human actions like car fires, or cigarette butts. We've seen it everywhere recently Brazil, Australia, United States, Canada, etc., destroying plant, human, and animal life, property damage, and contributing to global warming through the high amount of CO2 produced. But thanks to AI, we may be able to spot these fires much sooner and take action sooner. Here's how artificial intelligence can be used to reduce fire detection time from an average of 40 minutes to less than five minutes!

In brief OpenAI has disbanded its AI robotics team and is no longer trying to apply machine learning to physical machines. Wojciech Zaremba, co-founder of OpenAI, who led the robotics group confirmed that the company recently broke up the team to focus working on more promising areas of artificial general intelligence research. "Here's a reveal ... as of recently we changed the focus at OpenAI, and I actually disbanded the robotics team," he said during an episode of the Weights & Biases podcast. Zaremba said a lack of training data was holding the robotics research back: there wasn't enough information on hand to teach the systems to the level of intelligence desired. "From the perspective of what we want to achieve, which is to build AGI, I think there was actually some components missing," he added.

I asked OpenAI's now-famous algorithm GPT-3 to write me a story. GPT-3 is likely one of the most powerful natural language processing (NLP) algorithms in the world. It can be used for a wide range of tasks, such as summarizing articles, powering video game dialogue, and even writing programming code.

Natural language understanding: Smart assistants and conversational interactive voice response systems leverage deep learning to comprehend natural human speech that enables users to make requests. For example, instead of a customer calling a traditional call center, they could ask their smart speaker to initiate a product return. From there, the voice assistant can open the company's voice app and guide the customer through the return process. Deep learning and natural language understanding software recognizes that requests such as, "I'd like to make a return" and "Can I send my purchase back?" mean the same thing, so consumers don't have to remember certain phrases or words to initiate a task.

PyTorch uses the autograd package for automatic differentiation. For a tensor y, we can calculate the gradient with respect to input with two methods. After we do the .backward(), PyTorch generates a Dynamic Computation Graph when the forward function of network is called. We borrow a toy example from here.

Soft robotics is a thriving branch of robotics which takes inspiration from nature and uses affordable flexible materials to design adaptable non-rigid robots. However, their flexible behavior makes these robots hard to model, which is essential for a precise actuation and for optimal control. For system modelling, learning-based approaches have demonstrated good results, yet they fail to consider the physical structure underlying the system as an inductive prior. In this work, we take inspiration from sensorimotor learning, and apply a Graph Neural Network to the problem of modelling a non-rigid kinematic chain (i.e. a robotic soft hand) taking advantage of two key properties: 1) the system is compositional, that is, it is composed of simple interacting parts connected by edges, 2) it is order invariant, i.e. only the structure of the system is relevant for predicting future trajectories. We denote our model as the 'Sensorimotor Graph' since it learns the system connectivity from observation and uses it for dynamics prediction. We validate our model in different scenarios and show that it outperforms the non-structured baselines in dynamics prediction while being more robust to configurational variations, tracking errors or node failures.

Tractably modelling distributions over manifolds has long been an important goal in the natural sciences. Recent work has focused on developing general machine learning models to learn such distributions. However, for many applications these distributions must respect manifold symmetries -- a trait which most previous models disregard. In this paper, we lay the theoretical foundations for learning symmetry-invariant distributions on arbitrary manifolds via equivariant manifold flows. We demonstrate the utility of our approach by using it to learn gauge invariant densities over $SU(n)$ in the context of quantum field theory.

Bayesian neural network (BNN) allows for uncertainty quantification in prediction, offering an advantage over regular neural networks that has not been explored in the differential privacy (DP) framework. We fill this important gap by leveraging recent development in Bayesian deep learning and privacy accounting to offer a more precise analysis of the trade-off between privacy and accuracy in BNN. We propose three DP-BNNs that characterize the weight uncertainty for the same network architecture in distinct ways, namely DP-SGLD (via the noisy gradient method), DP-BBP (via changing the parameters of interest) and DP-MC Dropout (via the model architecture). Interestingly, we show a new equivalence between DP-SGD and DP-SGLD, implying that some non-Bayesian DP training naturally allows for uncertainty quantification. However, the hyperparameters such as learning rate and batch size, can have different or even opposite effects in DP-SGD and DP-SGLD. Extensive experiments are conducted to compare DP-BNNs, in terms of privacy guarantee, prediction accuracy, uncertainty quantification, calibration, computation speed, and generalizability to network architecture. As a result, we observe a new tradeoff between the privacy and the reliability. When compared to non-DP and non-Bayesian approaches, DP-SGLD is remarkably accurate under strong privacy guarantee, demonstrating the great potential of DP-BNN in real-world tasks.

Machine learning methods, such as deep learning, show promising results in the medical domain. However, the lack of interpretability of these algorithms may hinder their applicability to medical decision support systems. This paper studies an interpretable deep learning technique, called SincNet. SincNet is a convolutional neural network that efficiently learns customized band-pass filters through trainable sinc-functions. In this study, we use SincNet to analyze the neural activity of individuals with Autism Spectrum Disorder (ASD), who experience characteristic differences in neural oscillatory activity. In particular, we propose a novel SincNet-based neural network for detecting emotions in ASD patients using EEG signals. The learned filters can be easily inspected to detect which part of the EEG spectrum is used for predicting emotions. We found that our system automatically learns the high-$\alpha$ (9-13 Hz) and $\beta$ (13-30 Hz) band suppression often present in individuals with ASD. This result is consistent with recent neuroscience studies on emotion recognition, which found an association between these band suppressions and the behavioral deficits observed in individuals with ASD. The improved interpretability of SincNet is achieved without sacrificing performance in emotion recognition.