NeurIPS 2022

BYOL-Explore: Exploration by bootstrapped prediction

We present BYOL-Explore, a conceptually simple yet general approach for curiosity-driven exploration in visually-complex environments. BYOL-Explore learns a world representation, the world dynamics, and an exploration policy all-together by optimizing a single prediction loss in the latent space with no additional auxiliary objective. We show that BYOL-Explore is effective in DM-HARD-8, a challenging partially-observable continuous-action hard-exploration benchmark with visually-rich 3-D environments. On this benchmark, we solve the majority of the tasks purely through augmenting the extrinsic reward with BYOL-Explore s intrinsic reward, whereas prior work could only get off the ground with human demonstrations. As further evidence of the generality of BYOL-Explore, we show that it achieves superhuman performance on the ten hardest exploration games in Atari while having a much simpler design than other competitive agents.

Santiago Cuervo on Twitter

“A bit overdue, but still glad to introduce our work to wrap up the year: Variable-rate hierarchical CPC leads to acoustic unit discovery in speech ↕️ ⌛💬🧠 presented at #NeurIPS2022. #Speech #AI #DL #RL #SignalProcessing #SelfSupervised 📜: https://t.co/ISHU2jF9eX 🧵👇(1/n)”

This is your brain. This is your brain on code

Researchers from MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL) found that the Multiple Demand and Language brain systems encode specific code properties and uniquely align with machine-learned representations of code.

Censored Quantile Regression Neural Networks for Distribution-Free Survival Analysis

This paper considers doing quantile regression on censored data using neural networks (NNs). This adds to the survival analysis toolkit by allowing direct prediction of the target variable, along with a distribution-free characterisation of uncertainty, using a flexible function approximator. We begin by showing how an algorithm popular in linear models can be applied to NNs. However, the resulting procedure is inefficient, requiring sequential optimisation of an individual NN at each desired quantile. Our major contribution is a novel algorithm that simultaneously optimises a grid of quantiles output by a single NN. To offer theoretical insight into our algorithm, we show firstly that it can be interpreted as a form of expectation-maximisation, and secondly that it exhibits a desirable `self-correcting' property. Experimentally, the algorithm produces quantiles that are better calibrated than existing methods on 10 out of 12 real datasets.

MineDojo | Building Open-Ended Embodied Agents with Internet-Scale Knowledge