Reinforcement learning with human feedback for aligning large language models (LLMs) trains a reward model typically using ranking loss with comparison pairs.However, the training procedure suffers from an inherent problem: the uncontrolled scaling of reward scores during reinforcement learning due to the lack of constraints while training the reward model.This paper proposes a Prior Constraints-based Reward Model (namely PCRM) training method to mitigate this problem. PCRM incorporates prior constraints, specifically, length ratio and cosine similarity between outputs of each comparison pair, during reward model training to regulate optimization magnitude and control score margins. We comprehensively evaluate PCRM by examining its rank correlation with human preferences and its effectiveness in aligning LLMs via RL. Experimental results demonstrate that PCRM significantly improves alignment performance by effectively constraining reward score scaling. As another bonus, our method is easily integrated into arbitrary rank-based alignment methods, such as direct preference optimization, and can yield consistent improvement.
Understanding the mechanisms behind Vision Transformer (ViT), particularly its vulnerability to adversarial perturba tions, is crucial for addressing challenges in its real-world applications. Existing ViT adversarial attackers rely on la bels to calculate the gradient for perturbation, and exhibit low transferability to other structures and tasks. In this paper, we present a label-free white-box attack approach for ViT-based models that exhibits strong transferability to various black box models, including most ViT variants, CNNs, and MLPs, even for models developed for other modalities. Our inspira tion comes from the feature collapse phenomenon in ViTs, where the critical attention mechanism overly depends on the low-frequency component of features, causing the features in middle-to-end layers to become increasingly similar and eventually collapse. We propose the feature diversity attacker to naturally accelerate this process and achieve remarkable performance and transferability.
With the advancement of generative modeling techniques, synthetic human speech becomes increasingly indistinguishable from real, and tricky challenges are elicited for the audio deepfake detection (ADD) system. In this paper, we exploit audio features to improve the generalizability of ADD systems. Investigation of the ADD task performance is conducted over a broad range of audio features, including various handcrafted features and learning-based features. Experiments show that learning-based audio features pretrained on a large amount of data generalize better than hand-crafted features on out-of-domain scenarios. Subsequently, we further improve the generalizability of the ADD system using proposed multi-feature approaches to incorporate complimentary information from features of different views. The model trained on ASV2019 data achieves an equal error rate of 24.27\% on the In-the-Wild dataset.
While considerable progress has been made in achieving accurate lip synchronization for 3D speech-driven talking face generation, the task of incorporating expressive facial detail synthesis aligned with the speaker's speaking status remains challenging. Our goal is to directly leverage the inherent style information conveyed by human speech for generating an expressive talking face that aligns with the speaking status. In this paper, we propose AVI-Talking, an Audio-Visual Instruction system for expressive Talking face generation. This system harnesses the robust contextual reasoning and hallucination capability offered by Large Language Models (LLMs) to instruct the realistic synthesis of 3D talking faces. Instead of directly learning facial movements from human speech, our two-stage strategy involves the LLMs first comprehending audio information and generating instructions implying expressive facial details seamlessly corresponding to the speech. Subsequently, a diffusion-based generative network executes these instructions. This two-stage process, coupled with the incorporation of LLMs, enhances model interpretability and provides users with flexibility to comprehend instructions and specify desired operations or modifications. Extensive experiments showcase the effectiveness of our approach in producing vivid talking faces with expressive facial movements and consistent emotional status.
Image-to-image translation (I2I), and particularly its subfield of appearance transfer, which seeks to alter the visual appearance between images while maintaining structural coherence, presents formidable challenges. Despite significant advancements brought by diffusion models, achieving fine-grained transfer remains complex, particularly in terms of retaining detailed structural elements and ensuring information fidelity. This paper proposes an innovative framework designed to surmount these challenges by integrating various aspects of semantic matching, appearance transfer, and latent deviation. A pivotal aspect of our approach is the strategic use of the predicted $x_0$ space by diffusion models within the latent space of diffusion processes. This is identified as a crucial element for the precise and natural transfer of fine-grained details. Our framework exploits this space to accomplish semantic alignment between source and target images, facilitating mask-wise appearance transfer for improved feature acquisition. A significant advancement of our method is the seamless integration of these features into the latent space, enabling more nuanced latent deviations without necessitating extensive model retraining or fine-tuning. The effectiveness of our approach is demonstrated through extensive experiments, which showcase its ability to adeptly handle fine-grained appearance transfers across a wide range of categories and domains. We provide our code at https://github.com/babahui/Fine-grained-Appearance-Transfer
We present an unsupervised 3D shape co-segmentation method which learns a set of deformable part templates from a shape collection. To accommodate structural variations in the collection, our network composes each shape by a selected subset of template parts which are affine-transformed. To maximize the expressive power of the part templates, we introduce a per-part deformation network to enable the modeling of diverse parts with substantial geometry variations, while imposing constraints on the deformation capacity to ensure fidelity to the originally represented parts. We also propose a training scheme to effectively overcome local minima. Architecturally, our network is a branched autoencoder, with a CNN encoder taking a voxel shape as input and producing per-part transformation matrices, latent codes, and part existence scores, and the decoder outputting point occupancies to define the reconstruction loss. Our network, coined DAE-Net for Deforming Auto-Encoder, can achieve unsupervised 3D shape co-segmentation that yields fine-grained, compact, and meaningful parts that are consistent across diverse shapes. We conduct extensive experiments on the ShapeNet Part dataset, DFAUST, and an animal subset of Objaverse to show superior performance over prior methods.
Recent advances in 3D content creation mostly leverage optimization-based 3D generation via score distillation sampling (SDS). Though promising results have been exhibited, these methods often suffer from slow per-sample optimization, limiting their practical usage. In this paper, we propose DreamGaussian, a novel 3D content generation framework that achieves both efficiency and quality simultaneously. Our key insight is to design a generative 3D Gaussian Splatting model with companioned mesh extraction and texture refinement in UV space. In contrast to the occupancy pruning used in Neural Radiance Fields, we demonstrate that the progressive densification of 3D Gaussians converges significantly faster for 3D generative tasks. To further enhance the texture quality and facilitate downstream applications, we introduce an efficient algorithm to convert 3D Gaussians into textured meshes and apply a fine-tuning stage to refine the details. Extensive experiments demonstrate the superior efficiency and competitive generation quality of our proposed approach. Notably, DreamGaussian produces high-quality textured meshes in just 2 minutes from a single-view image, achieving approximately 10 times acceleration compared to existing methods.
In spite of the rapidly evolving landscape of text-to-image generation, the synthesis and manipulation of multiple entities while adhering to specific relational constraints pose enduring challenges. This paper introduces an innovative progressive synthesis and editing operation that systematically incorporates entities into the target image, ensuring their adherence to spatial and relational constraints at each sequential step. Our key insight stems from the observation that while a pre-trained text-to-image diffusion model adeptly handles one or two entities, it often falters when dealing with a greater number. To address this limitation, we propose harnessing the capabilities of a Large Language Model (LLM) to decompose intricate and protracted text descriptions into coherent directives adhering to stringent formats. To facilitate the execution of directives involving distinct semantic operations-namely insertion, editing, and erasing-we formulate the Stimulus, Response, and Fusion (SRF) framework. Within this framework, latent regions are gently stimulated in alignment with each operation, followed by the fusion of the responsive latent components to achieve cohesive entity manipulation. Our proposed framework yields notable advancements in object synthesis, particularly when confronted with intricate and lengthy textual inputs. Consequently, it establishes a new benchmark for text-to-image generation tasks, further elevating the field's performance standards.
Recent years have witnessed great progress in creating vivid audio-driven portraits from monocular videos. However, how to seamlessly adapt the created video avatars to other scenarios with different backgrounds and lighting conditions remains unsolved. On the other hand, existing relighting studies mostly rely on dynamically lighted or multi-view data, which are too expensive for creating video portraits. To bridge this gap, we propose ReliTalk, a novel framework for relightable audio-driven talking portrait generation from monocular videos. Our key insight is to decompose the portrait's reflectance from implicitly learned audio-driven facial normals and images. Specifically, we involve 3D facial priors derived from audio features to predict delicate normal maps through implicit functions. These initially predicted normals then take a crucial part in reflectance decomposition by dynamically estimating the lighting condition of the given video. Moreover, the stereoscopic face representation is refined using the identity-consistent loss under simulated multiple lighting conditions, addressing the ill-posed problem caused by limited views available from a single monocular video. Extensive experiments validate the superiority of our proposed framework on both real and synthetic datasets. Our code is released in https://github.com/arthur-qiu/ReliTalk.
Large language models achieve state-of-the-art performance on sequence generation evaluation, but typically have a large number of parameters. This is a computational challenge as presented by applying their evaluation capability at scale. To overcome the challenge, in this paper, we propose \textbf{ECT}, an \textbf{e}valuation \textbf{c}apability \textbf{t}ransfer method, to transfer the evaluation capability from LLMs to relatively lightweight language models. Based on the proposed ECT, we learn various evaluation models from ChatGPT, and employ them as reward models to improve sequence generation models via reinforcement learning and reranking approaches. Experimental results on machine translation, text style transfer, and summarization tasks demonstrate the effectiveness of our ECT. Notably, applying the learned evaluation models to sequence generation models results in better generated sequences as evaluated by commonly used metrics and ChatGPT.