Self-Correcting Decoding with Generative Feedback for Mitigating Hallucinations in Large Vision-Language Models

While recent Large Vision-Language Models (LVLMs) have shown remarkable performance in multi-modal tasks, they are prone to generating hallucinatory text responses that do not align with the given visual input, which restricts their practical applicability in real-world scenarios. In this work, inspired by the observation that the text-to-image generation process is the inverse of image-conditioned response generation in LVLMs, we explore the potential of leveraging text-to-image generative models to assist in mitigating hallucinations in LVLMs. We discover that generative models can offer valuable self-feedback for mitigating hallucinations at both the response and token levels. Building on this insight, we introduce self-correcting Decoding with Generative Feedback (DeGF), a novel training-free algorithm that incorporates feedback from text-to-image generative models into the decoding process to effectively mitigate hallucinations in LVLMs. Specifically, DeGF generates an image from the initial response produced by LVLMs, which acts as an auxiliary visual reference and provides self-feedback to verify and correct the initial response through complementary or contrastive decoding. Extensive experimental results validate the effectiveness of our approach in mitigating diverse types of hallucinations, consistently surpassing state-of-the-art methods across six benchmarks. Code is available at https://github.com/zhangce01/DeGF.

B3C: A Minimalist Approach to Offline Multi-Agent Reinforcement Learning

Overestimation arising from selecting unseen actions during policy evaluation is a major challenge in offline reinforcement learning (RL). A minimalist approach in the single-agent setting -- adding behavior cloning (BC) regularization to existing online RL algorithms -- has been shown to be effective; however, this approach is understudied in multi-agent settings. In particular, overestimation becomes worse in multi-agent settings due to the presence of multiple actions, resulting in the BC regularization-based approach easily suffering from either over-regularization or critic divergence. To address this, we propose a simple yet effective method, Behavior Cloning regularization with Critic Clipping (B3C), which clips the target critic value in policy evaluation based on the maximum return in the dataset and pushes the limit of the weight on the RL objective over BC regularization, thereby improving performance. Additionally, we leverage existing value factorization techniques, particularly non-linear factorization, which is understudied in offline settings. Integrated with non-linear value factorization, B3C outperforms state-of-the-art algorithms on various offline multi-agent benchmarks.

Aug3D: Augmenting large scale outdoor datasets for Generalizable Novel View Synthesis

Recent photorealistic Novel View Synthesis (NVS) advances have increasingly gained attention. However, these approaches remain constrained to small indoor scenes. While optimization-based NVS models have attempted to address this, generalizable feed-forward methods, offering significant advantages, remain underexplored. In this work, we train PixelNeRF, a feed-forward NVS model, on the large-scale UrbanScene3D dataset. We propose four training strategies to cluster and train on this dataset, highlighting that performance is hindered by limited view overlap. To address this, we introduce Aug3D, an augmentation technique that leverages reconstructed scenes using traditional Structure-from-Motion (SfM). Aug3D generates well-conditioned novel views through grid and semantic sampling to enhance feed-forward NVS model learning. Our experiments reveal that reducing the number of views per cluster from 20 to 10 improves PSNR by 10%, but the performance remains suboptimal. Aug3D further addresses this by combining the newly generated novel views with the original dataset, demonstrating its effectiveness in improving the model's ability to predict novel views.

PlanCritic: Formal Planning with Human Feedback

Real world planning problems are often too complex to be effectively tackled by a single unaided human. To alleviate this, some recent work has focused on developing a collaborative planning system to assist humans in complex domains, with bridging the gap between the system's problem representation and the real world being a key consideration. Transferring the speed and correctness formal planners provide to real-world planning problems is greatly complicated by the dynamic and online nature of such tasks. Formal specifications of task and environment dynamics frequently lack constraints on some behaviors or goal conditions relevant to the way a human operator prefers a plan to be carried out. While adding constraints to the representation with the objective of increasing its realism risks slowing down the planner, we posit that the same benefits can be realized without sacrificing speed by modeling this problem as an online preference learning task. As part of a broader cooperative planning system, we present a feedback-driven plan critic. This method makes use of reinforcement learning with human feedback in conjunction with a genetic algorithm to directly optimize a plan with respect to natural-language user preferences despite the non-differentiability of traditional planners. Directly optimizing the plan bridges the gap between research into more efficient planners and research into planning with language models by utilizing the convenience of natural language to guide the output of formal planners. We demonstrate the effectiveness of our plan critic at adhering to user preferences on a disaster recovery task, and observe improved performance compared to an llm-only neurosymbolic approach.

HiMemFormer: Hierarchical Memory-Aware Transformer for Multi-Agent Action Anticipation

Understanding and predicting human actions has been a long-standing challenge and is a crucial measure of perception in robotics AI. While significant progress has been made in anticipating the future actions of individual agents, prior work has largely overlooked a key aspect of real-world human activity -- interactions. To address this gap in human-like forecasting within multi-agent environments, we present the Hierarchical Memory-Aware Transformer (HiMemFormer), a transformer-based model for online multi-agent action anticipation. HiMemFormer integrates and distributes global memory that captures joint historical information across all agents through a transformer framework, with a hierarchical local memory decoder that interprets agent-specific features based on these global representations using a coarse-to-fine strategy. In contrast to previous approaches, HiMemFormer uniquely hierarchically applies the global context with agent-specific preferences to avoid noisy or redundant information in multi-agent action anticipation. Extensive experiments on various multi-agent scenarios demonstrate the significant performance of HiMemFormer, compared with other state-of-the-art methods.

Symbolic Graph Inference for Compound Scene Understanding

Scene understanding is a fundamental capability needed in many domains, ranging from question-answering to robotics. Unlike recent end-to-end approaches that must explicitly learn varying compositions of the same scene, our method reasons over their constituent objects and analyzes their arrangement to infer a scene's meaning. We propose a novel approach that reasons over a scene's scene- and knowledge-graph, capturing spatial information while being able to utilize general domain knowledge in a joint graph search. Empirically, we demonstrate the feasibility of our method on the ADE20K dataset and compare it to current scene understanding approaches.

DyPNIPP: Predicting Environment Dynamics for RL-based Robust Informative Path Planning

Informative path planning (IPP) is an important planning paradigm for various real-world robotic applications such as environment monitoring. IPP involves planning a path that can learn an accurate belief of the quantity of interest, while adhering to planning constraints. Traditional IPP methods typically require high computation time during execution, giving rise to reinforcement learning (RL) based IPP methods. However, the existing RL-based methods do not consider spatio-temporal environments which involve their own challenges due to variations in environment characteristics. In this paper, we propose DyPNIPP, a robust RL-based IPP framework, designed to operate effectively across spatio-temporal environments with varying dynamics. To achieve this, DyPNIPP incorporates domain randomization to train the agent across diverse environments and introduces a dynamics prediction model to capture and adapt the agent actions to specific environment dynamics. Our extensive experiments in a wildfire environment demonstrate that DyPNIPP outperforms existing RL-based IPP algorithms by significantly improving robustness and performing across diverse environment conditions.

Navigating Noisy Feedback: Enhancing Reinforcement Learning with Error-Prone Language Models

The correct specification of reward models is a well-known challenge in reinforcement learning. Hand-crafted reward functions often lead to inefficient or suboptimal policies and may not be aligned with user values. Reinforcement learning from human feedback is a successful technique that can mitigate such issues, however, the collection of human feedback can be laborious. Recent works have solicited feedback from pre-trained large language models rather than humans to reduce or eliminate human effort, however, these approaches yield poor performance in the presence of hallucination and other errors. This paper studies the advantages and limitations of reinforcement learning from large language model feedback and proposes a simple yet effective method for soliciting and applying feedback as a potential-based shaping function. We theoretically show that inconsistent rankings, which approximate ranking errors, lead to uninformative rewards with our approach. Our method empirically improves convergence speed and policy returns over commonly used baselines even with significant ranking errors, and eliminates the need for complex post-processing of reward functions.

Dual Prototype Evolving for Test-Time Generalization of Vision-Language Models

Test-time adaptation, which enables models to generalize to diverse data with unlabeled test samples, holds significant value in real-world scenarios. Recently, researchers have applied this setting to advanced pre-trained vision-language models (VLMs), developing approaches such as test-time prompt tuning to further extend their practical applicability. However, these methods typically focus solely on adapting VLMs from a single modality and fail to accumulate task-specific knowledge as more samples are processed. To address this, we introduce Dual Prototype Evolving (DPE), a novel test-time adaptation approach for VLMs that effectively accumulates task-specific knowledge from multi-modalities. Specifically, we create and evolve two sets of prototypes--textual and visual--to progressively capture more accurate multi-modal representations for target classes during test time. Moreover, to promote consistent multi-modal representations, we introduce and optimize learnable residuals for each test sample to align the prototypes from both modalities. Extensive experimental results on 15 benchmark datasets demonstrate that our proposed DPE consistently outperforms previous state-of-the-art methods while also exhibiting competitive computational efficiency. Code is available at https://github.com/zhangce01/DPE-CLIP.

OffRIPP: Offline RL-based Informative Path Planning

Informative path planning (IPP) is a crucial task in robotics, where agents must design paths to gather valuable information about a target environment while adhering to resource constraints. Reinforcement learning (RL) has been shown to be effective for IPP, however, it requires environment interactions, which are risky and expensive in practice. To address this problem, we propose an offline RL-based IPP framework that optimizes information gain without requiring real-time interaction during training, offering safety and cost-efficiency by avoiding interaction, as well as superior performance and fast computation during execution -- key advantages of RL. Our framework leverages batch-constrained reinforcement learning to mitigate extrapolation errors, enabling the agent to learn from pre-collected datasets generated by arbitrary algorithms. We validate the framework through extensive simulations and real-world experiments. The numerical results show that our framework outperforms the baselines, demonstrating the effectiveness of the proposed approach.