Several recent works demonstrate that transformers can implement algorithms like gradient
descent. By a careful construction of weights, these works show that multiple layers of …

Highlights•Neural network models of vision are dominated by feedforward architectures.•
Biological vision, by contrast, exhibits abundant recurrent processing.•The computational …

In this paper, we identify and characterize the emerging area of representation engineering
(RepE), an approach to enhancing the transparency of AI systems that draws on insights …

We analyze transformers from the perspective of iterative inference, seeking to understand
how model predictions are refined layer by layer. To do so, we train an affine probe for each …

Although deep learning has produced dazzling successes for applications of image, speech,
and video processing in the past few years, most trainings are with suboptimal hyper …

The representation of geometry in real-time 3D perception systems continues to be a critical
research issue. Dense maps capture complete surface shape and can be augmented with …

A large body of research in continual learning is devoted to overcoming the catastrophic
forgetting of neural networks by designing new algorithms that are robust to the distribution …

This paper is on highly accurate and highly efficient human pose estimation. Recent works
based on Fully Convolutional Networks (FCNs) have demonstrated excellent results for this …

Deep residual networks (ResNets) and their variants are widely used in many computer
vision applications and natural language processing tasks. However, the theoretical …

Deep convolutional artificial neural networks (ANNs) are the leading class of candidate
models of the mechanisms of visual processing in the primate ventral stream. While initially …