Foundation models have exhibited unprecedented capabilities across various domains and tasks. Models like CLIP bridge cross-modal representations, while text-to-image diffusion models excel in realistic image generation. While the complexity of these models makes retraining infeasible, their superior performance has driven research to explore how to efficiently use them for downstream tasks. Our work explores how to leverage these models for dense visual prediction tasks, specifically image segmentation. To avoid the annotation cost or training large diffusion models, we constrain our method to be zero-shot and training-free. Our pipeline, dubbed FreeSeg-Diff, uses open-source foundation models to perform open-vocabulary segmentation as follows: (a) retrieving image caption (via BLIP-2) and visual features (via Stable Diffusion), (b) clustering and binarizing features to form class-agnostic object masks, (c) mapping these masks to textual classes using CLIP with open vocabulary support, and (d) refining coarse masks. FreeSeg-Diff surpasses many training-based methods on Pascal VOC and COCO datasets and delivers competitive results against recent weakly-supervised segmentation approaches. We provide experiments demonstrating the superiority of diffusion model features over other pre-trained models. Project page: https://bcorrad.github.io/freesegdiff/.

The rapid growth of transformer-based models increases the concerns about their integrity and ownership insurance. Watermarking addresses this issue by embedding a unique identifier into the model, while preserving its performance. However, most existing approaches require to optimize the weights to imprint the watermark signal, which is not suitable at scale due to the computational cost. This paper explores watermarks with virtually no computational cost, applicable to a non-blind white-box setting (assuming access to both the original and watermarked networks). They generate functionally equivalent copies by leveraging the models’ invariance, via operations like dimension permutations or scaling/unscaling. This enables to watermark models without any change in their outputs and remains stealthy. Experiments demonstrate the effectiveness of the approach and its robustness against various model transformations (fine-tuning, quantization, pruning), making it a practical solution to protect the integrity of large models.

Denoising Diffusion Probabilistic Models (DDPMs) have recently achieved remarkable results in conditional and unconditional image generation. The pre-trained models can be adapted without further training to different downstream tasks, by guiding their iterative denoising process at inference time to satisfy additional constraints. For the specific task of image inpainting, the current guiding mechanism relies on copying-and-pasting the known regions from the input image at each denoising step. However, diffusion models are strongly conditioned by the initial random noise, and therefore struggle to harmonize predictions inside the inpainting mask with the real parts of the input image, often producing results with unnatural artifacts. Our method, dubbed GradPaint, steers the generation towards a globally coherent image. At each step in the denoising process, we leverage the model’s “denoised image estimation” by calculating a custom loss measuring its coherence with the masked input image. Our guiding mechanism uses the gradient obtained from backpropagating this loss through the diffusion model itself. GradPaint generalizes well to diffusion models trained on various datasets, improving upon current state-of-the-art supervised and unsupervised methods. Our code will be made available upon publication.

Large-scale text-to-image diffusion models have significantly improved the state of the art in generative image modeling and allow for an intuitive and powerful user interface to drive the image generation process. Expressing spatial constraints, e.g. to position specific objects in particular locations, is cumbersome using text; and current text-based image generation models are not able to accurately follow such instructions. In this paper we consider image generation from text associated with segments on the image canvas, which combines an intuitive natural language interface with precise spatial control over the generated content. We propose ZestGuide, a "zero-shot" segmentation guidance approach that can be plugged into pre-trained text-to-image diffusion models, and does not require any additional training. It leverages implicit segmentation maps that can be extracted from cross-attention layers, and uses them to align the generation with input masks. Our experimental results combine high image quality with accurate alignment of generated content with input segmentations, and improve over prior work both quantitatively and qualitatively, including methods that require training on images with corresponding segmentations. Compared to Paint with Words, the previous state-of-the art in image generation with zero-shot segmentation conditioning, we improve by 5 to 10 mIoU points on the COCO dataset with similar FID scores.

Generative image modeling enables a wide range of applications but raises ethical concerns about responsible deployment. We introduce an active content tracing method combining image watermarking and Latent Diffusion Models. The goal is for all generated images to conceal an invisible watermark allowing for future detection and/or identification. The method quickly fine-tunes the latent decoder of the image generator, conditioned on a binary signature. A pre-trained watermark extractor recovers the hidden signature from any generated image and a statistical test then determines whether it comes from the generative model. We evaluate the invisibility and robustness of the watermarks on a variety of generation tasks, showing that the Stable Signature is robust to image modifications. For instance, it detects the origin of an image generated from a text prompt, then cropped to keep 10% of the content, with 90+% accuracy at a false positive rate below 10−6.

Foundation models are first pre-trained on vast unsupervised datasets and then fine-tuned on labeled data. Reinforcement learning, notably from human feedback (RLHF), can further align the network with the intended usage. Yet the imperfections in the proxy reward may hinder the training and lead to suboptimal results; the diversity of objectives in real-world tasks and human opinions exacerbate the issue. This paper proposes embracing the heterogeneity of diverse rewards by following a multi-policy strategy. Rather than focusing on a single a priori reward, we aim for Pareto-optimal generalization across the entire space of preferences. To this end, we propose rewarded soup, first specializing multiple networks independently (one for each proxy reward) and then interpolating their weights linearly. This succeeds empirically because we show that the weights remain linearly connected when fine-tuned on diverse rewards from a shared pre-trained initialization. We demonstrate the effectiveness of our approach for text-to-text (summarization, Q&A, helpful assistant, review), text-image (image captioning, text-to-image generation, visual grounding), and control (locomotion) tasks. We hope to enhance the alignment of deep models, and how they interact with the world in all its diversity.

Deep generative models, like GANs, have considerably improved the state of the art in image synthesis, and are able to generate near photo-realistic images in structured domains such as human faces. Based on this success, recent work on image editing proceeds by projecting images to the GAN latent space and manipulating the latent vector. However, these approaches are limited in that only images from a narrow domain can be transformed, and with only a limited number of editing operations. We propose FlexIT, a novel method which can take any input image and a user-defined text instruction for editing. Our method achieves flexible and natural editing, pushing the limits of semantic image translation. First, FlexIT combines the input image and text into a single target point in the CLIP multimodal embedding space. Via the latent space of an autoencoder, we iteratively transform the input image toward the target point, ensuring coherence and quality with a variety of novel regularization terms. We propose an evaluation protocol for semantic image translation, and thoroughly evaluate our method on ImageNet. Code will be available at https://github.com/facebookresearch/SemanticImageTranslation/.

Deep network architectures struggle to continually learn new tasks without forgetting the previous tasks. A recent trend indicates that dynamic architectures based on an ex-pansion of the parameters can reduce catastrophic forget-ting efficiently in continual learning. However, existing approaches often require a task identifier at test-time, need complex tuning to balance the growing number of parameters, and barely share any information across tasks. As a result, they struggle to scale to a large number of tasks without significant overhead. In this paper, we propose a transformer architecture based on a dedicated encoder/decoder framework. Critically, the encoder and decoder are shared among all tasks. Through a dynamic expansion of special tokens, we specialize each forward of our decoder network on a task distribution. Our strategy scales to a large number of tasks while having neg-ligible memory and time overheads due to strict control of the expansion of the parameters. Moreover, this efficient strategy doesn't need any hyperparameter tuning to control the network's expansion. Our model reaches excellent results on CIFAR100 and state-of-the-art performances on the large-scale ImageNet100 and ImageNet100 while having fewer parameters than concurrent dynamic frameworks.11Code is released at https://github.com/arthurdouillard/dytox.

State-of-the-art vision and vision-and-language models rely on large-scale visio-linguistic pretraining for obtaining good performance on a variety of downstream tasks. Generally, such models are often either cross-modal (contrastive) or multi-modal (with earlier fusion) but not both; and they often only target specific modalities or tasks. A promising direction would be to use a single holistic universal model, as a “foundation”, that targets all modalities at once-a true vision and language foundation model should be good at vision tasks, language tasks, and cross- and multi-modal vision and language tasks. We introduce FLAVA as such a model and demonstrate impressive performance on a wide range of 35 tasks spanning these target modalities.

The vast expansion of protein sequence databases provides an opportunity for new protein design approaches which seek to learn the sequence-function relationship directly from natural sequence variation. Deep generative models trained on protein sequence data have been shown to learn biologically meaningful representations helpful for a variety of downstream tasks, but their potential for direct use in the design of novel proteins remains largely unexplored. Here we show that variational autoencoders trained on a dataset of almost 70000 luciferase-like oxidoreductases can be used to generate novel, functional variants of the luxA bacterial luciferase. We propose separate VAE models to work with aligned sequence input (MSA VAE) and raw sequence input (AR-VAE), and offer evidence that while both are able to reproduce patterns of amino acid usage characteristic of the family, the MSA VAE is better able to capture long-distance dependencies reflecting the influence of 3D structure. To confirm the practical utility of the models, we used them to generate variants of luxA whose luminescence activity was validated experimentally. We further showed that conditional variants of both models could be used to increase the solubility of luxA without disrupting function. Altogether 6/12 of the variants generated using the unconditional AR-VAE and 9/11 generated using the unconditional MSA VAE retained measurable luminescence, together with all 23 of the less distant variants generated by conditional versions of the models; the most distant functional variant contained 35 differences relative to the nearest training set sequence. These results demonstrate the feasibility of using deep generative models to explore the space of possible protein sequences and generate useful variants, providing a method complementary to rational design and directed evolution approaches.