Artificial Intelligence

Excessive Invariance Causes Adversarial Vulnerability

Uses bijective networks to identify large subspaces of invariance-based vulnerability and introduces the independence cross-entropy loss which partially alleviates it.

Scaling and Benchmarking Self-Supervised Visual Representation Learning

Demonstrates that scaling up self-supervised methods along data size, model capacity, and problem complexity enables them to match or surpass ImageNet supervised pre-training on a variety of tasks.

On The Power of Curriculum Learning in Training Deep Networks

Demonstrates the benefit of curriculum learning with different scoring and pacing functions on various small datasets.

Learning Transferable Visual Models From Natural Language Supervision

Applies task-agnostic, web-scale pre-training to computer vision using natural language supervision, enabling powerful zero-shot transfer to many datasets.

ConvNets and ImageNet Beyond Accuracy: Understanding Mistakes and Uncovering Biases

Demonstrates that providing explantions and model criticism can be useful tools to improve the reliability of ImageNet-trained CNNs for end-users.

Could a Neuroscientist Understand a Microprocessor?

Addresses a popular belief in neuroscience that the field is primarily data limited by using a microprocessor as a model organism and applying modern data analysis methods from neuroscience to understand its information processing, with generally poor results.

Embodied Intelligence via Learning and Evolution

Large-scale evolutionary simulations by DERL yield insights into how the interaction between learning, evolution, and environmental complexity can lead to morphological intelligence.

Training data-efficient image transformers & distillation through attention

Produces competitive convolution-free transformer, training only on ImageNet.

Intrinsically motivated learning of real-world sensorimotor skills with developmental constraints

The unlearnability, high-dimensionality, and unboundedness of the real world necessitates the integration of intrinsic motivation with other developmental constraints, such as sensorimotor primitives, task space representations, maturational mechanisms, and social guidance.

Learning the Predictability of the Future

Presents the idea of using hyperbolic embeddings for hierarchical representations and provides some experiments classifying action within a hierarchy of actions.