Alex Tamkin

TL;DR: The authors provides a thorough account of the opportunities and risks of foundation models, ranging from their capabilities (e.g., language, vision, robotics, reasoning, human interaction) and technical principles(e. g.g. model architectures, training procedures, data, systems, security, evaluation, theory) to their applications.

TL;DR: This technique reveals that in BERT, layers with high probing performance on downstream GLUE tasks are neither necessary nor sufficient for high accuracy on those tasks, and the benefit of using pretrained parameters for a layer varies dramatically with dataset size.

TL;DR: The design process of drone.io is described, a projected body-centric graphical user interface for human-drone interaction embedded on a drone to provide both input and output capabilities, and it is reported that people were able to use the interface with little prior training.

TL;DR: This work proposes viewmaker networks: generative models that learn to produce input-dependent views for contrastive learning, and demonstrates that learned views are a promising way to reduce the amount of expertise and effort needed for unsupervised learning, potentially extending its benefits to a much wider set of domains.

TL;DR: This paper presents the first algorithm for sample-efficient learning of CVaR-optimal policies in Markov decision processes based on the optimism in the face of uncertainty principle by relying on a novel optimistic version of the distributional Bellman operator that moves probability mass from the lower to the upper tail of the return distribution.