From Lists to Emojis: How Format Bias Affects Model Alignment

In this paper, we study format biases in reinforcement learning from human feedback (RLHF). We observe that many widely-used preference models, including human evaluators, GPT-4, and top-ranking models on the RewardBench benchmark, exhibit strong biases towards specific format patterns, such as lists, links, bold text, and emojis. Furthermore, large language models (LLMs) can exploit these biases to achieve higher rankings on popular benchmarks like AlpacaEval and LMSYS Chatbot Arena. One notable example of this is verbosity bias, where current preference models favor longer responses that appear more comprehensive, even when their quality is equal to or lower than shorter, competing responses. However, format biases beyond verbosity remain largely underexplored in the literature. In this work, we extend the study of biases in preference learning beyond the commonly recognized length bias, offering a comprehensive analysis of a wider range of format biases. Additionally, we show that with a small amount of biased data (less than 1%), we can inject significant bias into the reward model. Moreover, these format biases can also be easily exploited by downstream alignment algorithms, such as best-of-n sampling and online iterative DPO, as it is usually easier to manipulate the format than to improve the quality of responses. Our findings emphasize the need to disentangle format and content both for designing alignment algorithms and evaluating models.

Revisiting Adaptive Cellular Recognition Under Domain Shifts: A Contextual Correspondence View

Cellular nuclei recognition serves as a fundamental and essential step in the workflow of digital pathology. However, with disparate source organs and staining procedures among histology image clusters, the scanned tiles inherently conform to a non-uniform data distribution, which induces deteriorated promises for general cross-cohort usages. Despite the latest efforts leveraging domain adaptation to mitigate distributional discrepancy, those methods are subjected to modeling the morphological characteristics of each cell individually, disregarding the hierarchical latent structure and intrinsic contextual correspondences across the tumor micro-environment. In this work, we identify the importance of implicit correspondences across biological contexts for exploiting domain-invariant pathological composition and thereby propose to exploit the dependence over various biological structures for domain adaptive cellular recognition. We discover those high-level correspondences via unsupervised contextual modeling and use them as bridges to facilitate adaptation over diverse organs and stains. In addition, to further exploit the rich spatial contexts embedded amongst nuclear communities, we propose self-adaptive dynamic distillation to secure instance-aware trade-offs across different model constituents. The proposed method is extensively evaluated on a broad spectrum of cross-domain settings under miscellaneous data distribution shifts and outperforms the state-of-the-art methods by a substantial margin. Code is available at https://github.com/camwew/CellularRecognition_DA_CC.

MLPHand: Real Time Multi-View 3D Hand Mesh Reconstruction via MLP Modeling

Multi-view hand mesh reconstruction is a critical task for applications in virtual reality and human-computer interaction, but it remains a formidable challenge. Although existing multi-view hand reconstruction methods achieve remarkable accuracy, they typically come with an intensive computational burden that hinders real-time inference. To this end, we propose MLPHand, a novel method designed for real-time multi-view single hand reconstruction. MLP Hand consists of two primary modules: (1) a lightweight MLP-based Skeleton2Mesh model that efficiently recovers hand meshes from hand skeletons, and (2) a multi-view geometry feature fusion prediction module that enhances the Skeleton2Mesh model with detailed geometric information from multiple views. Experiments on three widely used datasets demonstrate that MLPHand can reduce computational complexity by 90% while achieving comparable reconstruction accuracy to existing state-of-the-art baselines.

Not All Prompts Are Made Equal: Prompt-based Pruning of Text-to-Image Diffusion Models

Text-to-image (T2I) diffusion models have demonstrated impressive image generation capabilities. Still, their computational intensity prohibits resource-constrained organizations from deploying T2I models after fine-tuning them on their internal target data. While pruning techniques offer a potential solution to reduce the computational burden of T2I models, static pruning methods use the same pruned model for all input prompts, overlooking the varying capacity requirements of different prompts. Dynamic pruning addresses this issue by utilizing a separate sub-network for each prompt, but it prevents batch parallelism on GPUs. To overcome these limitations, we introduce Adaptive Prompt-Tailored Pruning (APTP), a novel prompt-based pruning method designed for T2I diffusion models. Central to our approach is a prompt router model, which learns to determine the required capacity for an input text prompt and routes it to an architecture code, given a total desired compute budget for prompts. Each architecture code represents a specialized model tailored to the prompts assigned to it, and the number of codes is a hyperparameter. We train the prompt router and architecture codes using contrastive learning, ensuring that similar prompts are mapped to nearby codes. Further, we employ optimal transport to prevent the codes from collapsing into a single one. We demonstrate APTP's effectiveness by pruning Stable Diffusion (SD) V2.1 using CC3M and COCO as target datasets. APTP outperforms the single-model pruning baselines in terms of FID, CLIP, and CMMD scores. Our analysis of the clusters learned by APTP reveals they are semantically meaningful. We also show that APTP can automatically discover previously empirically found challenging prompts for SD, e.g., prompts for generating text images, assigning them to higher capacity codes.

From Pixels to Prose: A Large Dataset of Dense Image Captions

Training large vision-language models requires extensive, high-quality image-text pairs. Existing web-scraped datasets, however, are noisy and lack detailed image descriptions. To bridge this gap, we introduce PixelProse, a comprehensive dataset of over 16M (million) synthetically generated captions, leveraging cutting-edge vision-language models for detailed and accurate descriptions. To ensure data integrity, we rigorously analyze our dataset for problematic content, including child sexual abuse material (CSAM), personally identifiable information (PII), and toxicity. We also provide valuable metadata such as watermark presence and aesthetic scores, aiding in further dataset filtering. We hope PixelProse will be a valuable resource for future vision-language research. PixelProse is available at https://huggingface.co/datasets/tomg-group-umd/pixelprose

OPTune: Efficient Online Preference Tuning

Reinforcement learning with human feedback~(RLHF) is critical for aligning Large Language Models (LLMs) with human preference. Compared to the widely studied offline version of RLHF, \emph{e.g.} direct preference optimization (DPO), recent works have shown that the online variants achieve even better alignment. However, online alignment requires on-the-fly generation of new training data, which is costly, hard to parallelize, and suffers from varying quality and utility. In this paper, we propose a more efficient data exploration strategy for online preference tuning (OPTune), which does not rely on human-curated or pre-collected teacher responses but dynamically samples informative responses for on-policy preference alignment. During data generation, OPTune only selects prompts whose (re)generated responses can potentially provide more informative and higher-quality training signals than the existing responses. In the training objective, OPTune reweights each generated response (pair) by its utility in improving the alignment so that learning can be focused on the most helpful samples. Throughout our evaluations, OPTune'd LLMs maintain the instruction-following benefits provided by standard preference tuning whilst enjoying 1.27-1.56x faster training speed due to the efficient data exploration strategy.

Event3DGS: Event-based 3D Gaussian Splatting for Fast Egomotion

The recent emergence of 3D Gaussian splatting (3DGS) leverages the advantage of explicit point-based representations, which significantly improves the rendering speed and quality of novel-view synthesis. However, 3D radiance field rendering in environments with high-dynamic motion or challenging illumination condition remains problematic in real-world robotic tasks. The reason is that fast egomotion is prevalent real-world robotic tasks, which induces motion blur, leading to inaccuracies and artifacts in the reconstructed structure. To alleviate this problem, we propose Event3DGS, the first method that learns Gaussian Splatting solely from raw event streams. By exploiting the high temporal resolution of event cameras and explicit point-based representation, Event3DGS can reconstruct high-fidelity 3D structures solely from the event streams under fast egomotion. Our sparsity-aware sampling and progressive training approaches allow for better reconstruction quality and consistency. To further enhance the fidelity of appearance, we explicitly incorporate the motion blur formation process into a differentiable rasterizer, which is used with a limited set of blurred RGB images to refine the appearance. Extensive experiments on multiple datasets validate the superior rendering quality of Event3DGS compared with existing approaches, with over 95% lower training time and faster rendering speed in orders of magnitude.

Distortion-free Watermarks are not Truly Distortion-free under Watermark Key Collisions

Language model (LM) watermarking techniques inject a statistical signal into LM-generated content by substituting the random sampling process with pseudo-random sampling, using watermark keys as the random seed. Among these statistical watermarking approaches, distortion-free watermarks are particularly crucial because they embed watermarks into LM-generated content without compromising generation quality. However, one notable limitation of pseudo-random sampling compared to true-random sampling is that, under the same watermark keys (i.e., key collision), the results of pseudo-random sampling exhibit correlations. This limitation could potentially undermine the distortion-free property. Our studies reveal that key collisions are inevitable due to the limited availability of watermark keys, and existing distortion-free watermarks exhibit a significant distribution bias toward the original LM distribution in the presence of key collisions. Moreover, achieving a perfect distortion-free watermark is impossible as no statistical signal can be embedded under key collisions. To reduce the distribution bias caused by key collisions, we introduce a new family of distortion-free watermarks--beta-watermark. Experimental results support that the beta-watermark can effectively reduce the distribution bias under key collisions.

Interpretable Spatio-Temporal Embedding for Brain Structural-Effective Network with Ordinary Differential Equation

The MRI-derived brain network serves as a pivotal instrument in elucidating both the structural and functional aspects of the brain, encompassing the ramifications of diseases and developmental processes. However, prevailing methodologies, often focusing on synchronous BOLD signals from functional MRI (fMRI), may not capture directional influences among brain regions and rarely tackle temporal functional dynamics. In this study, we first construct the brain-effective network via the dynamic causal model. Subsequently, we introduce an interpretable graph learning framework termed Spatio-Temporal Embedding ODE (STE-ODE). This framework incorporates specifically designed directed node embedding layers, aiming at capturing the dynamic interplay between structural and effective networks via an ordinary differential equation (ODE) model, which characterizes spatial-temporal brain dynamics. Our framework is validated on several clinical phenotype prediction tasks using two independent publicly available datasets (HCP and OASIS). The experimental results clearly demonstrate the advantages of our model compared to several state-of-the-art methods.

Paintings and Drawings Aesthetics Assessment with Rich Attributes for Various Artistic Categories

Image aesthetic evaluation is a highly prominent research domain in the field of computer vision. In recent years, there has been a proliferation of datasets and corresponding evaluation methodologies for assessing the aesthetic quality of photographic works, leading to the establishment of a relatively mature research environment. However, in contrast to the extensive research in photographic aesthetics, the field of aesthetic evaluation for paintings and Drawings has seen limited attention until the introduction of the BAID dataset in March 2023. This dataset solely comprises overall scores for high-quality artistic images. Our research marks the pioneering introduction of a multi-attribute, multi-category dataset specifically tailored to the field of painting: Aesthetics of Paintings and Drawings Dataset (APDD). The construction of APDD received active participation from 28 professional artists worldwide, along with dozens of students specializing in the field of art. This dataset encompasses 24 distinct artistic categories and 10 different aesthetic attributes. Each image in APDD has been evaluated by six professionally trained experts in the field of art, including assessments for both total aesthetic scores and aesthetic attribute scores. The final APDD dataset comprises a total of 4985 images, with an annotation count exceeding 31100 entries. Concurrently, we propose an innovative approach: Art Assessment Network for Specific Painting Styles (AANSPS), designed for the assessment of aesthetic attributes in mixed-attribute art datasets. Through this research, our goal is to catalyze advancements in the field of aesthetic evaluation for paintings and drawings, while enriching the available resources and methodologies for its further development and application.