NeurIPS 2024 论文列表

When making decisions under uncertainty, individuals often deviate from rational behavior, which can be evaluated across three dimensions: risk preference, probability weighting, and loss aversion. Given the widespread use of large language models (LLMs) in supporting decision-making processes, it is crucial to assess whether their behavior aligns with human norms and ethical expectations or exhibits potential biases. Although several empirical studies have investigated the rationality and social behavior performance of LLMs, their internal decision-making tendencies and capabilities remain inadequately understood. This paper proposes a framework, grounded in behavioral economics theories, to evaluate the decision-making behaviors of LLMs. With a multiple-choice-list experiment, we initially estimate the degree of risk preference, probability weighting, and loss aversion in a context-free setting for three commercial LLMs: ChatGPT-4.0-Turbo, Claude-3-Opus, and Gemini-1.0-pro. Our results reveal that LLMs generally exhibit patterns similar to humans, such as risk aversion and loss aversion, with a tendency to overweight small probabilities, but there are significant variations in the degree to which these behaviors are expressed across different LLMs. Further, we explore their behavior when embedded with socio-demographic features of human beings, uncovering significant disparities across various demographic characteristics.

Existing ML benchmarks lack the depth and diversity of annotations needed for evaluating models on business process management (BPM) tasks. BPM is the practice of documenting, measuring, improving, and automating enterprise workflows. However, research has focused almost exclusively on one task -- full end-to-end automation using agents based on multimodal foundation models (FMs) like GPT-4. This focus on automation ignores the reality of how most BPM tools are applied today -- simply documenting the relevant workflow takes 60% of the time of the typical process optimization project. To address this gap we present WONDERBREAD, the first benchmark for evaluating multimodal FMs on BPM tasks beyond automation. Our contributions are: (1) a dataset containing 2928 documented workflow demonstrations; (2) 6 novel BPM tasks sourced from real-world applications ranging from workflow documentation to knowledge transfer to process improvement; and (3) an automated evaluation harness. Our benchmark shows that while state-of-the-art FMs can automatically generate documentation (e.g. recalling 88% of the steps taken in a video demonstration of a workflow), they struggle to re-apply that knowledge towards finer-grained validation of workflow completion (F1 < 0.3). We hope WONDERBREAD encourages the development of more "human-centered" AI tooling for enterprise applications and furthers the exploration of multimodal FMs for the broader universe of BPM tasks. We publish our dataset and experiments here: https://github.com/HazyResearch/wonderbread

Vision-and-Language Navigation (VLN) requires an agent to dynamically explore environments following natural language. The VLN agent, closely integrated into daily lives, poses a substantial threat to the security of privacy and property upon the occurrence of malicious behavior. However, this serious issue has long been overlooked. In this paper, we pioneer the exploration of an object-aware backdoored VLN, achieved by implanting object-aware backdoors during the training phase. Tailored to the unique VLN nature of cross-modality and continuous decision-making, we propose a novel backdoored VLN paradigm: IPR Backdoor. This enables the agent to act in abnormal behavior once encountering the object triggers during language-guided navigation in unseen environments, thereby executing an attack on the target scene. Experiments demonstrate the effectiveness of our method in both physical and digital spaces across different VLN agents, as well as its robustness to various visual and textual variations. Additionally, our method also well ensures navigation performance in normal scenarios with remarkable stealthiness.

Data curation is the problem of how to collect and organize samples into a dataset that supports efficient learning. Despite the centrality of the task, little work has been devoted towards a large-scale, systematic comparison of various curation methods. In this work, we take steps towards a formal evaluation of data curation strategies and introduce SELECT, the first large-scale benchmark of curation strategies for image classification. In order to generate baseline methods for the SELECT benchmark, we create a new dataset, ImageNet++, which constitutes the largest superset of ImageNet-1K to date. Our dataset extends ImageNet with 5 new training-data shifts, each approximately the size of ImageNet-1K, and each assembled using a distinct curation strategy. We evaluate our data curation baselines in two ways: (i) using each training-data shift to train identical image classification models from scratch (ii) using it to inspect a fixed pretrained self-supervised representation. Our findings show interesting trends, particularly pertaining to recent methods for data curation such as synthetic data generation and lookup based on CLIP embeddings. We show that although these strategies are highly competitive for certain tasks, the curation strategy used to assemble the original ImageNet-1K dataset remains the gold standard. We anticipate that our benchmark can illuminate the path for new methods to further reduce the gap. We release our checkpoints, code, documentation, and a link to our dataset at https://github.com/jimmyxu123/SELECT.

Continual learning (CL) aims to help deep neural networks to learn new knowledge while retaining what has been learned. Owing to their powerful generalizability, pre-trained vision-language models such as Contrastive Language-Image Pre-training (CLIP) have lately gained traction as practical CL candidates. However, the domain mismatch between the pre-training and the downstream CL tasks calls for finetuning of the CLIP on the latter. The deterministic nature of the existing finetuning methods makes them overlook the many possible interactions across the modalities and deems them unsafe for high-risk tasks requiring reliable uncertainty estimation. To address these, our work proposes **C**ontinual **L**e**A**rning with **P**robabilistic finetuning (CLAP) - a probabilistic modeling framework over visual-guided text features per task, thus providing more calibrated CL finetuning. Unlike recent data-hungry anti-forgetting CL techniques, CLAP alleviates forgetting by exploiting the rich pre-trained knowledge of CLIP for weight initialization and distribution regularization of task-specific parameters. Cooperating with the diverse range of existing prompting methods, CLAP can surpass the predominant deterministic finetuning approaches for CL with CLIP. We conclude with out-of-the-box applications of superior uncertainty estimation abilities of CLAP including novel data detection and exemplar selection within the existing CL setups. Our code is available at https://github.com/srvCodes/clap4clip.

Vision-language models have shown impressive progress in recent years. However, existing models are largely limited to turn-based interactions, where each turn must be stepped (i.e., prompted) by the user. Open-ended, asynchronous interactions, where an AI model may proactively deliver timely responses or feedback based on the unfolding situation in real-time, are an open challenge. In this work, we present the QEVD benchmark and dataset, which explores human-AI interaction in the challenging, yet controlled, real-world domain of fitness coaching – a task which intrinsically requires monitoring live user activity and providing immediate feedback. The benchmark requires vision-language models to recognize complex human actions, identify possible mistakes, and provide appropriate feedback in real-time. Our experiments reveal the limitations of existing state-of-the-art vision-language models for such asynchronous situated interactions. Motivated by this, we propose a simple end-to-end streaming baseline that can respond asynchronously to human actions with appropriate feedback at the appropriate time.

Data selection has emerged as a core issue for large-scale visual-language model pretaining (e.g., CLIP), particularly with noisy web-curated datasets. Three main data selection approaches are: (1) leveraging external non-CLIP models to aid data selection, (2) training new CLIP-style embedding models that are more effective at selecting high-quality data than the original OpenAI CLIP model, and (3) designing better metrics or strategies universally applicable to any CLIP embedding without requiring specific model properties (e.g., CLIPScore is one popular metric). While the first two approaches have been extensively studied, the third remains under-explored. In this paper, we advance the third approach by proposing two new methods. Firstly, instead of classical CLIP scores that only consider the alignment between two modalities from a single sample, we introduce $\textbf{negCLIPLoss}$, a method inspired by CLIP training loss that adds the alignment between one sample and its contrastive pairs as an extra normalization term to CLIPScore for better quality measurement. Secondly, when downstream tasks are known, we propose a new norm-based metric, $\textbf{NormSim}$, to measure the similarity between pretraining data and target data. We test our methods on the data selection benchmark, DataComp [Gadre et al., 2023]. Compared to the best baseline using only OpenAI's CLIP-L/14, our methods achieve a 5.3\% improvement on ImageNet-1k and a 2.8\% improvement on 38 downstream evaluation tasks. Moreover, both $\textbf{negCLIPLoss}$ and $\textbf{NormSim}$ are compatible with existing techniques. By combining our methods with the current best methods DFN [Fang et al., 2023] and HYPE [Kim et al., 2024], we can boost average performance on downstream tasks by 0.9\%, achieving a new state-of-the-art on the DataComp-medium benchmark.

A straightforward pipeline for zero-shot out-of-distribution (OOD) detection involves selecting potential OOD labels from an extensive semantic pool and then leveraging a pre-trained vision-language model to perform classification on both in-distribution (ID) and OOD labels. In this paper, we theorize that enhancing performance requires expanding the semantic pool, while increasing the expected probability of selected OOD labels being activated by OOD samples, and ensuring low mutual dependence among the activations of these OOD labels. A natural expansion manner is to adopt a larger lexicon; however, the inevitable introduction of numerous synonyms and uncommon words fails to meet the above requirements, indicating that viable expansion manners move beyond merely selecting words from a lexicon. Since OOD detection aims to correctly classify input images into ID/OOD class groups, we can "make up" OOD label candidates which are not standard class names but beneficial for the process. Observing that the original semantic pool is comprised of unmodified specific class names, we correspondingly construct a conjugated semantic pool (CSP) consisting of modified superclass names, each serving as a cluster center for samples sharing similar properties across different categories. Consistent with our established theory, expanding OOD label candidates with the CSP satisfies the requirements and outperforms existing works by 7.89% in FPR95. Codes are available in https://github.com/MengyuanChen21/NeurIPS2024-CSP.

Recently, large language models (LLMs) have achieved superior performance, empowering the development of large multimodal agents (LMAs). An LMA is anticipated to execute practical tasks requires various capabilities including multimodal perception, interaction, reasoning, and decision making. However, existing benchmarks are limited in assessing compositional skills and actions demanded by practical scenarios, where they primarily focused on single tasks and static scenarios. To bridge this gap, we introduce WhodunitBench, a benchmark rooted from murder mystery games, where players are required to utilize the aforementioned skills to achieve their objective (i.e., identifying the `murderer' or hiding themselves), providing a simulated dynamic environment for evaluating LMAs. Specifically, WhodunitBench includes two evaluation modes. The first mode, the arena-style evaluation, is constructed from 50 meticulously curated scripts featuring clear reasoning clues and distinct murderers; The second mode, the chain of evaluation, consists of over 3000 curated multiple-choice questions and open-ended questions, aiming to assess every facet of the murder mystery games for LMAs. Experiments show that although current LMAs show acceptable performance in basic perceptual tasks, they are insufficiently equipped for complex multi-agent collaboration and multi-step reasoning tasks. Furthermore, the full application of the theory of mind to complete games in a manner akin to human behavior remains a significant challenge. We hope this work can illuminate the path forward, providing a solid foundation for the future development of LMAs. Our WhodunitBench is open-source and accessible at: https://github.com/ jun0wanan/WhodunitBench-Murder_Mystery_Games

Benchmarks for large multimodal language models (MLMs) now serve to simultaneously assess the general capabilities of models instead of evaluating for a specific capability. As a result, when a developer wants to identify which models to use for their application, they are overwhelmed by the number of benchmarks and remain uncertain about which benchmark's results are most reflective of their specific use case. This paper introduces Task-Me-Anything, a benchmark generation engine which produces a benchmark tailored to a user's needs. Task-Me-Anything maintains an extendable taxonomy of visual assets and can programmatically generate a vast number of task instances. Additionally, it algorithmically addresses user queries regarding MLM performance efficiently within a computational budget. It contains 113K images, 10K videos, 2K 3D object assets, over 365 object categories, 655 attributes, and 335 relationships. It can generate 500M image/video question-answering pairs, which focus on evaluating MLM perceptual capabilities. Task-Me-Anything reveals critical insights: open-source MLMs excel in object and attribute recognition but lack spatial and temporal understanding; each model exhibits unique strengths and weaknesses; larger models generally perform better, though exceptions exist; and GPT4O demonstrates challenges in recognizing rotating/moving objects and distinguishing colors.

Recently, the diffusion-based generative paradigm has achieved impressive general image generation capabilities with text prompts due to its accurate distribution modeling and stable training process. However, generating diverse remote sensing (RS) images that are tremendously different from general images in terms of scale and perspective remains a formidable challenge due to the lack of a comprehensive remote sensing image generation dataset with various modalities, ground sample distances (GSD), and scenes. In this paper, we propose a Multi-modal, Multi-GSD, Multi-scene Remote Sensing (MMM-RS) dataset and benchmark for text-to-image generation in diverse remote sensing scenarios. Specifically, we first collect nine publicly available RS datasets and conduct standardization for all samples. To bridge RS images to textual semantic information, we utilize a large-scale pretrained vision-language model to automatically output text prompts and perform hand-crafted rectification, resulting in information-rich text-image pairs (including multi-modal images). In particular, we design some methods to obtain the images with different GSD and various environments (e.g., low-light, foggy) in a single sample. With extensive manual screening and refining annotations, we ultimately obtain a MMM-RS dataset that comprises approximately 2.1 million text-image pairs. Extensive experimental results verify that our proposed MMM-RS dataset allows off-the-shelf diffusion models to generate diverse RS images across various modalities, scenes, weather conditions, and GSD. The dataset is available at https://github.com/ljl5261/MMM-RS.

Accurate emotion perception is crucial for various applications, including human-computer interaction, education, and counseling. However, traditional single-modality approaches often fail to capture the complexity of real-world emotional expressions, which are inherently multimodal. Moreover, existing Multimodal Large Language Models (MLLMs) face challenges in integrating audio and recognizing subtle facial micro-expressions. To address this, we introduce the MERR dataset, containing 28,618 coarse-grained and 4,487 fine-grained annotated samples across diverse emotional categories. This dataset enables models to learn from varied scenarios and generalize to real-world applications. Furthermore, we propose Emotion-LLaMA, a model that seamlessly integrates audio, visual, and textual inputs through emotion-specific encoders. By aligning features into a shared space and employing a modified LLaMA model with instruction tuning, Emotion-LLaMA significantly enhances both emotional recognition and reasoning capabilities. Extensive evaluations show Emotion-LLaMA outperforms other MLLMs, achieving top scores in Clue Overlap (7.83) and Label Overlap (6.25) on EMER, an F1 score of 0.9036 on MER2023-SEMI challenge, and the highest UAR (45.59) and WAR (59.37) in zero-shot evaluations on DFEW dataset.

Image retargeting is the task of adjusting the aspect ratio of images to suit different display devices or presentation environments. However, existing retargeting methods often struggle to balance the preservation of key semantics and image quality, resulting in either deformation or loss of important objects, or the introduction of local artifacts such as discontinuous pixels and inconsistent regenerated content. To address these issues, we propose a content-aware retargeting method called PruneRepaint. It incorporates semantic importance for each pixel to guide the identification of regions that need to be pruned or preserved in order to maintain key semantics. Additionally, we introduce an adaptive repainting module that selects image regions for repainting based on the distribution of pruned pixels and the proportion between foreground size and target aspect ratio, thus achieving local smoothness after pruning. By focusing on the content and structure of the foreground, our PruneRepaint approach adaptively avoids key content loss and deformation, while effectively mitigating artifacts with local repainting. We conduct experiments on the public RetargetMe benchmark and demonstrate through objective experimental results and subjective user studies that our method outperforms previous approaches in terms of preserving semantics and aesthetics, as well as better generalization across diverse aspect ratios. Codes will be available at https://github.com/fhshen2022/PruneRepaint.

Current multimodal and multitask foundation models, like 4M or UnifiedIO, show promising results. However, their out-of-the-box abilities to accept diverse inputs and perform diverse tasks are limited by the (usually small) number of modalities and tasks they are trained on. In this paper, we develop a single any-to-any model trained on tens of highly diverse modalities and by performing co-training on large-scale multimodal datasets and text corpora. This includes training on images and text along with several semantic and geometric modalities, feature maps from recent state of the art models like DINOv2 and ImageBind, pseudo labels of specialist models like SAM and 4DHumans, and a range of new modalities that allow for novel ways to interact with the model and steer the generation, for example, image metadata or color palettes. A crucial step in this process is performing discrete tokenization on various modalities, whether they are image-like, neural network feature maps, vectors, structured data like instance segmentation or human poses, or data that can be represented as text. Through this, we show the possibility of training one model to solve at least 3x more tasks/modalities than existing models and doing so without a loss in performance. In addition, this enables more fine-grained and controllable multimodal generation capabilities and allows studying the distillation of models trained on diverse data and objectives into one unified model. We scale the training to a three billion parameter and different datasets. The multimodal models and training code are open sourced at https://4m.epfl.ch/.

Recent advances in self-supervised models for natural language, vision, and protein sequences have inspired the development of large genomic DNA language models (DNALMs). These models aim to learn generalizable representations of diverse DNA elements, potentially enabling various genomic prediction, interpretation and design tasks. Despite their potential, existing benchmarks do not adequately assess the capabilities of DNALMs on key downstream applications involving an important class of non-coding DNA elements critical for regulating gene activity. In this study, we introduce DART-Eval, a suite of representative benchmarks specifically focused on regulatory DNA to evaluate model performance across zero-shot, probed, and fine-tuned scenarios against contemporary ab initio models as baselines. Our benchmarks target biologically meaningful downstream tasks such as functional sequence feature discovery, predicting cell-type specific regulatory activity, and counterfactual prediction of the impacts of genetic variants. We find that current DNALMs exhibit inconsistent performance and do not offer compelling gains over alternative baseline models for most tasks, while requiring significantly more computational resources. We discuss potentially promising modeling, data curation, and evaluation strategies for the next generation of DNALMs. Our code is available at https://github.com/kundajelab/DART-Eval

Visible-Infrared Person Re-identification (VI-ReID) is a challenging cross-modal retrieval task due to significant modality differences, primarily resulting from the absence of color information in the infrared modality. The development of large foundation models like Large Language Models (LLMs) and Vision Language Models (VLMs) motivates us to explore a feasible solution to empower VI-ReID with off-the-shelf large foundation models. To this end, we propose a novel Text-enhanced VI-ReID framework driven by Large Foundation Models (TVI-LFM). The core idea is to enrich the representation of the infrared modality with textual descriptions automatically generated by VLMs. Specifically, we incorporate a pre-trained VLM to extract textual features from texts generated by VLM and augmented by LLM, and incrementally fine-tune the text encoder to minimize the domain gap between generated texts and original visual modalities. Meanwhile, to enhance the infrared modality with extracted textual representations, we leverage modality alignment capabilities of VLMs and VLM-generated feature-level filters. This enables the text model to learn complementary features from the infrared modality, ensuring the semantic structural consistency between the fusion modality and the visible modality. Furthermore, we introduce modality joint learning to align features across all modalities, ensuring that textual features maintain stable semantic representation of overall pedestrian appearance during complementary information learning. Additionally, a modality ensemble retrieval strategy is proposed to leverage complementary strengths of each query modality to improve retrieval effectiveness and robustness. Extensive experiments on three expanded VI-ReID datasets demonstrate that our method significantly improves the retrieval performance, paving the way for the utilization of large foundation models in downstream multi-modal retrieval tasks.

Vision Language Models (VLMs) demonstrate remarkable proficiency in addressing a wide array of visual questions, which requires strong perception and reasoning faculties. Assessing these two competencies independently is crucial for model refinement, despite the inherent difficulty due to the intertwined nature of seeing and reasoning in existing VLMs. To tackle this issue, we present Prism, an innovative framework designed to disentangle the perception and reasoning processes involved in visual question solving. Prism comprises two distinct stages: a perception stage that utilizes a VLM to extract and articulate visual information in textual form, and a reasoning stage that formulates responses based on the extracted visual information using a Large Language Model (LLM). This modular design enables the systematic comparison and assessment of both proprietary and open-source VLM for their perception and reasoning strengths. Our analytical framework provides several valuable insights, underscoring Prism's potential as a cost-effective solution for vision-language tasks. By combining a streamlined VLM focused on perception with a powerful LLM tailored for reasoning, Prism achieves superior results in general vision-language tasks while substantially cutting down on training and operational expenses. Quantitative evaluations show that Prism, when configured with a vanilla 2B LLaVA and freely accessible GPT-3.5, delivers performance on par with VLMs $10 \times$ larger on the rigorous multimodal benchmark MMStar.

Understanding the semantics of individual regions or patches of unconstrained images, such as open-world object detection, remains a critical yet challenging task in computer vision. Building on the success of powerful image-level vision-language (ViL) foundation models like CLIP, recent efforts have sought to harness their capabilities by either training a contrastive model from scratch with an extensive collection of region-label pairs or aligning the outputs of a detection model with image-level representations of region proposals. Despite notable progress, these approaches are plagued by computationally intensive training requirements, susceptibility to data noise, and deficiency in contextual information. To address these limitations, we explore the synergistic potential of off-the-shelf foundation models, leveraging their respective strengths in localization and semantics. We introduce a novel, generic, and efficient architecture, named RegionSpot, designed to integrate position-aware localization knowledge from a localization foundation model (e.g., SAM) with semantic information from a ViL model (e.g., CLIP). To fully exploit pretrained knowledge while minimizing training overhead, we keep both foundation models frozen, focusing optimization efforts solely on a lightweight attention-based knowledge integration module. Extensive experiments in open-world object recognition show that our RegionSpot achieves significant performance gain over prior alternatives, along with substantial computational savings (e.g., training our model with 3 million data in a single day using 8 V100 GPUs). RegionSpot outperforms GLIP-L by 2.9 in mAP on LVIS val set, with an even larger margin of 13.1 AP for more challenging and rare categories, and a 2.5 AP increase on ODinW. Furthermore, it exceeds GroundingDINO-L by 11.0 AP for rare categories on the LVIS minival set.

Physical Human-Scene Interaction (HSI) plays a crucial role in numerous applications. However, existing HSI techniques are limited to specific object dynamics and privileged information, which prevents the development of more comprehensive applications. To address this limitation, we introduce HumanVLA for general object rearrangement directed by practical vision and language. A teacher-student framework is utilized to develop HumanVLA. A state-based teacher policy is trained first using goal-conditioned reinforcement learning and adversarial motion prior. Then, it is distilled into a vision-language-action model via behavior cloning. We propose several key insights to facilitate the large-scale learning process. To support general object rearrangement by physical humanoid, we introduce a novel Human-in-the-Room dataset encompassing various rearrangement tasks. Through extensive experiments and analysis, we demonstrate the effectiveness of our approach.