Beyond Surface Statistics: Scene Representations in a Latent Diffusion Model · HackerPapers discussions

Beyond Surface Statistics: Scene Representations in a Latent Diffusion Model — mentioned by l33tman

Fri, 26 Jan 2024 01:29:17 GMT

Please see https://arxiv.org/abs/2306.05720 Beyond Surface Statistics: Scene Representations in a Latent Diffusion Model "... In this work, we investigate a basic interpretability question: does an LDM create and use an internal representation of simple scene geometry? Using linear probes, we find evidence that the internal activations of the LDM encode linear representations of both 3D depth data and a salient-object / background distinction. These representations appear surprisingly early in the denoising process−well before a human can easily make sense of the noisy images. ..."

Paper and discussion history. A mention is not an endorsement.

Beyond Surface Statistics: Scene Representations in a Latent Diffusion Model — mentioned by l33tman

Tue, 28 Nov 2023 16:21:52 GMT

Certainly: https://arxiv.org/abs/2306.05720 It's a very interesting paper. "Even when trained purely on images without explicit depth information, they typically output coherent pictures of 3D scenes. In this work, we investigate a basic interpretability question: does an LDM create and use an internal representation of simple scene geometry? Using linear probes, we find evidence that the internal activations of the LDM encode linear representations of both 3D depth data and a salient-object / background distinction. These representations appear surprisingly early in the denoising…

Paper and discussion history. A mention is not an endorsement.

Beyond Surface Statistics: Scene Representations in a Latent Diffusion Model — HN submission

Mon, 21 Aug 2023 13:31:55 GMT

Scene Representations in a Latent Diffusion Model

2 points; 1 discussion reply on HN. Paper and source history. A submission is not an endorsement.