Publications

How do diffusion generative models convert pure noise into meaningful images? In a variety of pretrained diffusion models (including conditional latent space models like Stable Diffusion), we observe that the reverse diffusion process that underlies image...

We developed an analytical framework for understanding how the learned distribution evolves during diffusion model training. Leveraging the Gaussian equivalence principle, we derived exact solutions for the gradient-flow dynamics of weights in one- or two...

Visual neurons respond across a vast landscape of images, comprising objects, textures, and places. Natural images can be parameterized using deep generative networks, raising the question of whether latent factors learned by some networks control images...

Humans excel at discovering regular structures from limited samples and applying inferred rules to novel settings. We investigate whether modern generative models can similarly learn underlying rules from finite samples and perform reasoning through...

By learning the gradient of smoothed data distributions, diffusion models can iteratively generate samples from complex distributions. The learned score function enables their generalization capabilities, but how the learned score relates to the score of...

Score-based models have achieved remarkable results in the generative modeling of many domains. By learning the gradient of smoothed data distribution, they can iteratively generate samples from complex distribution e.g. natural images.
However, is there...

How does feature learning happen during the training of a neural network? We developed an accelerated pipeline to synthesize maximally activating images ("prototypes") for hidden units in a parallel fashion. Through this, we were able to perform feature...

The natural world features high-dimensional retinal inputs to the visual system. In contrast, to study vision in the lab, only sparsely sampled image sets are used. Given this immense natural image manifold, what is a principled way to sample and...

Recently, optimization has become an emerging tool for neuroscientists to study neural code. In the visual system, neurons respond to images with graded and noisy responses. Image patterns eliciting highest responses are diagnostic of the coding content...

A goal in visual neuroscience is to explain how neurons respond to natural scenes. However, neurons are generally tested using simpler stimuli, often because they can be transformed smoothly, allowing the measurement of tuning functions (i.e., response...

Visual representations can be defined as the activations of neuronal populations in response to images. The activation of a neuron as a function over all image space has been described as a "tuning landscape". As a function over a high-dimensional space...