Subhash Kak
The paper analyzes the problem of judgments or preferences subsequent to initial analysis by autonomous agents in a hierarchical system where the higher level agents does not have access to group size information. We propose methods that reduce instances of preference reversal of the kind encountered in Simpson's paradox.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
继续探索
Hierarchical DQN (h-DQN) is a two-level architecture of feedforward neural networks where the meta level selects goals and the lower level takes actions to achieve the goals. We show tasks that cannot be solved by h-DQN, exemplifying the limitation of this type of hierarchical framework (HF). We describe the recurrent hierarchical framework (RHF), generalizing architectures that use a recurrent neural network at the meta level. We analyze the expressiveness of HF and RHF using context-sensitive grammars. We show that RHF is more expressive than HF. We perform experiments comparing an implement…
SheetCompass: Hierarchical Relation Graphs for Agentic Spreadsheet Reasoning
针对LLM处理电子表格时丢失多维结构语义的问题，提出SheetCompass框架，通过显式建模表内外结构关系并结合内存机制，提升智能体对复杂工作簿的推理能力。
Self-supervised Hierarchical Visual Reasoning with World Model
提出ResDreamer，一种分层世界模型，通过自监督方式学习残差表示，实现高效视觉推理，在3D开放环境中达到最先进的样本和参数效率。
ARISE: Agent Reasoning with Intrinsic Skill Evolution in Hierarchical Reinforcement Learning
ARISE提出一种分层强化学习框架，通过共享策略管理高阶技能和生成低阶响应，提升数学推理能力，实验显示其在多个基准测试中优于现有方法。
HeaRT: A Hierarchical Circuit Reasoning Tree-Based Agentic Framework for AMS Design Optimization
HeaRT提出了一种分层电路推理树的代理框架，通过提升F1(subcircuits)和F1(loops)指标，实现更高效的AMS设计优化，且在不同架构上表现出更好的适应性和收敛速度。
ReasoningLens: Hierarchical Visualization and Diagnostic Auditing for Large Reasoning Models
提出ReasoningLens框架，通过分层结构、智能审计和系统分析，将长思维链转化为可操作洞察，用于解释、调试和优化大型推理模型。
Hierarchical Budget Policy Optimization for Adaptive Reasoning
提出分层预算策略优化（HBPO），通过预算受限层级和差异化奖励，在保持探索多样性的同时，使模型自适应调整推理深度，平均令牌使用减少60.6%，准确率提升3.14%。
Reasoning about the behaviour of physical objects is a key capability of agents operating in physical worlds. Humans are very experienced in physical reasoning while it remains a major challenge for AI. To facilitate research addressing this problem, several benchmarks have been proposed recently. However, these benchmarks do not enable us to measure an agent's granular physical reasoning capabilities when solving a complex reasoning task. In this paper, we propose a new benchmark for physical reasoning that allows us to test individual physical reasoning capabilities. Inspired by how humans a…
Missing Premise exacerbates Overthinking: Are Reasoning Models losing Critical Thinking Skill?
本文发现推理型大语言模型在缺失前提问题上会出现 MiP-Overthinking 现象，响应冗长且无效，并揭示当前训练方案缺乏对高效思考的鼓励，为缓解过度思考问题提供了新见解。
PhotoCraft: Agentic Reasoning with Hierarchical Self-Evolving Memory for Deep Image Search
提出PhotoCraft，一种无需训练的分层记忆系统，通过工作、情景和语义记忆增强多模态大语言模型，实现深度图像搜索中的多步推理和知识迁移，在DISBench上提升检索性能达18.5%。
Neurosymbolic Reasoning with Incremental Knowledge for Sample Efficient Hierarchical Reinforcement Learning
本研究提出结合增量知识（InK）的神经符号分层强化学习，通过符号高层组件执行符号规划、低层神经模块学习运动原语，在导航任务中显著提升了样本效率，还开发了信念世界树搜索方法。
MemDreamer: Decoupling Perception and Reasoning for Long Video Understanding via Hierarchical Graph Memory and Agentic Retrieval Mechanism
提出MemDreamer框架，通过分层图记忆和智能体检索机制解耦感知与推理，将长视频理解转化为智能体探索过程，在四个基准上达到SOTA，推理上下文窗口仅占全量2%且准确率提升12.5点。
Homer: Understanding Long-form Videos with Hierarchical Memory and Agentic Reasoning
研究长视频理解问题，提出荷马分层在线记忆探索与推理框架，其记忆反映视频多尺度结构，智能推理器按人类方式探索记忆，通过多轮检索组合答案，性能优于此前最佳方法。
CHARMS: A Cognitive Hierarchical Agent for Reasoning and Motion Stylization in Autonomous Driving
CHARMS通过强化学习预训练和监督微调pipeline，实现人类般的决策行为和交互真实性，同时利用泊松认知层次理论生成多样化驾驶场景。
The broader goal of our research is to formulate answers to why and how questions with respect to knowledge bases, such as AURA. One issue we face when reasoning with many available knowledge bases is that at times needed information is missing. Examples of this include partially missing information about next sub-event, first sub-event, last sub-event, result of an event, input to an event, destination of an event, and raw material involved in an event. In many cases one can recover part of the missing knowledge through reasoning. In this paper we give a formal definition about how such missi…
Reverse Thinking Enhances Missing Information Detection in Large Language Models
本文提出逆向思维框架，通过引导LLM进行逆向推理来提升缺失信息检测的准确性与鲁棒性。
Visual Attention Reasoning via Hierarchical Search and Self-Verification
本文提出通过分层搜索与自验证的视觉注意力推理框架，有效提升多模态大语言模型的视觉定位和推理能力，显著降低幻觉发生率。
HiCrew: Hierarchical Reasoning for Long-Form Video Understanding via Question-Aware Multi-Agent Collaboration
本文提出HiCrew框架，通过混合树结构、问题感知描述生成和动态规划层，解决长视频中时空冗余和叙事依赖问题，提升时间与因果推理能力。
H2Table: Hierarchical Hypergraph-Enhanced Large Language Models for Complex Table Reasoning
针对大语言模型处理表格时忽略其二维与分层结构的问题，提出H2Table框架，通过分层嵌套超图及定制超图编码器等，在HiTab数据集嵌套深度4的复杂表格任务上较最优基线提升22.88%。
What's Missing in Screen-to-Action? Towards a UI-in-the-Loop Paradigm for Multimodal GUI Reasoning
提出UI-in-the-Loop (UILoop) 范式，通过循环的屏幕-UI元素-动作过程，让多模态大模型显式学习UI元素的定位、语义和用法，实现可解释推理，并在UI理解任务上达到最优。
Deconstructing Spatial Complexity: Hierarchical Decomposition for LLM Spatial Reasoning
提出一种层次任务分解方法，结合MCTS引导的组相对策略优化（M-GRPO），通过改进中间状态选择和规划能力，显著提升LLM在导航、规划和策略游戏等空间任务中的表现。
DHCG: Dynamic Construction of Hierarchical Collaboration Graphs for LLM-Based Multi-Agent Reasoning
DHCG将多智能体推理建模为部分可观测马尔可夫决策过程，动态构建分层协作图，通过规划器、执行器和生成器协作，在代码生成等基准上平均性能领先，较单智能体提升13.06点。
Hierarchical AI-Meteorologist: LLM-Agent System for Multi-Scale and Explainable Weather Forecast Reporting
本文提出分层人工智能气象学家，通过多尺度推理和关键词生成提升天气预报的可解释性和鲁棒性。
HiCoTraj:Zero-Shot Demographic Reasoning via Hierarchical Chain-of-Thought Prompting from Trajectory
针对轨迹数据人口统计推断依赖标注数据且泛化性差的问题，提出HiCoTraj框架，利用大语言模型零样本能力，将轨迹转为自然语言并通过层次化思维链推理，在真实数据上取得竞争性性能。
This paper presents a new technique for the design of approximate reasoning based controllers for dynamic physical systems with interacting goals. In this approach, goals are achieved based on a hierarchy defined by a control knowledge base and remain highly interactive during the execution of the control task. The approach has been implemented in a rule-based computer program which is used in conjunction with a prototype hardware system to solve the cart-pole balancing problem in real-time. It provides a complementary approach to the conventional analytical control methodology, and is of subs…
Hierarchically structured agent plans are important for efficient planning and acting, and they also serve (among other things) to produce "richer" classical plans, composed not just of a sequence of primitive actions, but also "abstract" ones representing the supplied hierarchies. A crucial step for this and other approaches is deriving precondition and effect "summaries" from a given plan hierarchy. This paper provides mechanisms to do this for more pragmatic and conventional hierarchies than in the past. To this end, we formally define the notion of a precondition and an effect for a hierar…
HyperTree Planning: Enhancing LLM Reasoning via Hierarchical Thinking
提出 HyperTree Planning (HTP) 推理范式，通过构建超树结构规划大纲并引入自主规划框架，使 LLM 能够分层思考并有效处理复杂规划任务，在 TravelPlanner 基准上取得 SOTA 准确率。
DocTrace: Towards Traceable Long Document VQA via Hierarchical Evidence Graph Reasoning
本文提出DocTrace分层框架，将长文档视觉问答建模为显式证据图推理问题，经两阶段训练优化后，在三个基准上优于现有模型，且推理可追溯。
Divide-Then-Rule: A Cluster-Driven Hierarchical Interpolator for Attribute-Missing Graphs
针对属性缺失图的深度图聚类难题，提出分而治之的DTRGC方法，通过动态簇感知传播、层次插补和跳级增强，显著提升聚类性能。
This paper presents an efficient adaptation and application of the Dempster-Shafer theory of evidence, one that can be used effectively in a massively parallel hierarchical system for visual pattern perception. It describes the techniques used, and shows in an extended example how they serve to improve the system's performance as it applies a multiple-level set of processes.
MapAgent: A Hierarchical Agent for Geospatial Reasoning with Dynamic Map Tool Integration
MapAgent提出分层多智能体框架，通过高层规划器分解任务并专用地图工具智能体并行编排API，在地理空间基准上超越现有方法，实现高效动态地图集成推理。
Hierarchical Attacks for Multi-Modal Multi-Agent Reasoning
本文提出HAM³框架，通过感知、通信、推理三层分层攻击，评估多智能体系统在不同推理范式下的攻击成功率，揭示多模态多智能体系统的安全漏洞。
Disrupting Hierarchical Reasoning: Adversarial Protection for Geographic Privacy in Multimodal Reasoning Models
本文提出ReasonBreak框架，通过概念感知扰动破坏多模态推理模型中的层次推理，利用GeoPrivacy-6K数据集验证其在提升地理隐私保护效果上的优势。
TableVision: A Large-Scale Benchmark for Spatially Grounded Reasoning over Complex Hierarchical Tables
本文提出TableVision基准，通过量化分析发现复杂表格推理中的感知瓶颈，引入轨迹感知的分层任务分类，结合确定性接地流程生成6799条高保真推理轨迹，提升多模态大语言模型的空间推理能力。
Hierarchical Latent Reasoning for LLM-based Recommendation
本文针对基于LLM的推荐系统提出分层潜在推理框架HiLaR，通过层感知强化优化提升推荐性能，在四个Amazon基准数据集上优于多种主流基线。
Evaluating Interactive Reasoning in Large Language Models: A Hierarchical Benchmark with Executable Games
提出一个多轮交互式推理评估框架，将推理视为主动证据获取和信念更新，通过474个可执行游戏基准测试大语言模型在成功率、交互效率、上下文鲁棒性和元认知适应方面的表现。
Cross-modal retrieval between videos and texts has attracted growing attentions due to the rapid emergence of videos on the web. The current dominant approach for this problem is to learn a joint embedding space to measure cross-modal similarities. However, simple joint embeddings are insufficient to represent complicated visual and textual details, such as scenes, objects, actions and their compositions. To improve fine-grained video-text retrieval, we propose a Hierarchical Graph Reasoning (HGR) model, which decomposes video-text matching into global-to-local levels. To be specific, the mode…
Numerical reasoning over hybrid data containing both textual and tabular content (e.g., financial reports) has recently attracted much attention in the NLP community. However, existing question answering (QA) benchmarks over hybrid data only include a single flat table in each document and thus lack examples of multi-step numerical reasoning across multiple hierarchical tables. To facilitate data analytical progress, we construct a new large-scale benchmark, MultiHiertt, with QA pairs over Multi Hierarchical Tabular and Textual data. MultiHiertt is built from a wealth of financial reports and …
The ability to reason with multiple hierarchical structures is an attractive and desirable property of sequential inductive biases for natural language processing. Do the state-of-the-art Transformers and LSTM architectures implicitly encode for these biases? To answer this, we propose ORCHARD, a diagnostic dataset for systematically evaluating hierarchical reasoning in state-of-the-art neural sequence models. While there have been prior evaluation frameworks such as ListOps or Logical Inference, our work presents a novel and more natural setting where our models learn to reason with multiple …
While discriminative classifiers often yield strong predictive performance, missing feature values at prediction time can still be a challenge. Classifiers may not behave as expected under certain ways of substituting the missing values, since they inherently make assumptions about the data distribution they were trained on. In this paper, we propose a novel framework that classifies examples with missing features by computing the expected prediction with respect to a feature distribution. Moreover, we use geometric programming to learn a naive Bayes distribution that embeds a given logistic r…
HiBench: Benchmarking LLMs Capability on Hierarchical Structure Reasoning
本文提出首个层次结构推理基准框架HiBench，包含6个场景、30个任务及39,519个查询，通过5个能力维度评估20个LLM，发现其在复杂和隐式层次结构上的不足，并构建指令数据集使Llama-3.1-8B和Qwen2.5-7B性能平均提升88.84%和31.38%。
CardioCoT: Hierarchical Reasoning for Multimodal Survival Analysis
针对急性心肌梗死患者术后MACE复发风险预测，提出CardioCoT两阶段分层推理框架，结合证据增强自精炼机制与影像数据，提升可解释性与预测性能。
THOR: A Theta-Gamma Hierarchical Oscillatory Reasoning Framework for Multi-hop QA
针对多跳问答中注意力衰减和错误积累问题，提出受大脑启发的θ-γ分层振荡推理框架THOR，经实验验证，该框架能提高答案准确性和鲁棒性，减轻相关限制。
SHAPE: Stage-aware Hierarchical Advantage via Potential Estimation for LLM Reasoning
SHAPE通过潜在估计引入分层信用分配机制，提升大语言模型推理的效率与准确性，实验显示在数学推理任务中平均准确率提升3%，token消耗减少30%。
MV-STRIDE: Enabling MLLMs to Master Multi-View Spatial Reasoning via Hierarchical Capability Modeling
针对多模态大模型多视角空间推理瓶颈，提出MV-STRIDE分层数据集与多阶段训练框架，显式建模能力依赖，在MMSI-Bench等基准上取得最优性能。
Hierarchical Reasoning Models: Perspectives and Misconceptions
本文回顾了分层推理模型，探讨其设计选择与变体，澄清误解，指出其在潜在空间循环推理中展现的潜力与待研究之处。
Are Your Reasoning Models Reasoning or Guessing? A Mechanistic Analysis of Hierarchical Reasoning Models
研究揭示分层推理模型在简单谜题中易失效，存在猜测动态和多固定点问题，提出数据增强、输入扰动和模型自举策略提升Sudoku-Extreme准确率至96.9%。
Knowledge Pyramid: A Novel Hierarchical Reasoning Structure for Generalized Knowledge Augmentation and Inference
本文针对知识图谱推理中低层知识导致泛化不足的问题，提出知识金字塔框架，通过从低层知识提取高层知识并在多层层次化知识图谱中进行推理，在医学数据集上验证了其能提升推理性能与泛化能力，尤其在少样本场景下准确率显著提高。
HERO: Hierarchical Evidential Reasoning Optimization for Radiology Report Generation via Reason-then-Summarize
本文提出了一种自洽的强化学习框架，用于生成可信的放射学报告，通过优化生成过程和减少幻觉，提升了临床效果。
Hierarchical Reasoning Model
针对大模型推理脆弱、数据需求大和延迟高的问题，受大脑层级处理启发，提出仅2700万参数、无需预训练和思维链数据的层级推理模型HRM，在数独、迷宫及ARC基准上性能卓越。
Abductive Inference in Retrieval-Augmented Language Models: Generating and Validating Missing Premises
本文提出将溯因推理集成到检索增强语言模型中，通过检测证据不足、生成并验证缺失前提，在多个基准上提升了答案准确性与推理忠实度。
Zero-Shot Human Mobility Forecasting via Large Language Model with Hierarchical Reasoning
本文提出ZHMF框架，将零样本人类移动预测转化为自然语言问答，通过语义增强检索、分层反思机制及活动规划与位置选择协同建模，有效捕捉长期意图与短期偏好，并在标准数据集上优于现有模型。
Think Hierarchically, Act Dynamically: Hierarchical Multi-modal Fusion and Reasoning for Vision-and-Language Navigation
本文提出多层级融合与推理架构（MFRA），通过分层融合视觉、语言与历史特征并动态推理，在REVERIE、R2R和SOON基准上超越现有方法，提升视觉语言导航决策准确性。
Hierarchical Mutual Information Analysis: Towards Multi-view Clustering in The Wild
本文提出一种深度多视图聚类框架，通过层次互信息最大化融合双重预测填补缺失视图与对比重建实现类别级对齐，首次分别处理视图缺失和未对齐问题，并在公开数据集上显著超越现有方法。
ReasonFlux: Hierarchical LLM Reasoning via Scaling Thought Templates
ReasonFlux通过约500个高层思维模板库、模板轨迹上的分层强化学习及推理时自适应模板扩展，仅用8块GPU训练32B模型，在MATH和AIME基准上超越o1-preview与DeepSeek-V3，实现最先进的数学推理能力。
HPP: Hierarchical Programmatic Probing for Long Video Understanding by Decoupling Perception and Reasoning
提出HPP框架，通过将长视频理解重构为分层视频的迭代程序化探索，解耦语义感知与高阶时序推理，在LongVideoBench等基准上取得显著提升。
Towards Hierarchical Multi-Step Reward Models for Enhanced Reasoning in Large Language Models
本文提出分层奖励模型和轻量数据增强策略，解决大语言模型推理中奖励黑客问题，提升多步推理评估的稳定性和泛化能力。
CosmicFish-HRM: Adaptive Reasoning via Hierarchical Recurrent Mechanisms in Compact Language Models
提出一种紧凑语言模型CosmicFish-HRM，通过层次推理模块动态分配推理深度，在保持较小参数量的同时实现适应性推理。
HMARS: A Hierarchical Multi-Agent Memory System for Long-Context Reasoning
提出HMARS分层多智能体记忆系统，通过子智能体、中层智能体和前沿模型的分层结构管理长上下文，在长文档和多轮记忆任务中优于检索、重排序、全上下文、基于图和智能体基线。
LocationAgent: A Hierarchical Agent for Image Geolocation via Decoupling Strategy and Evidence from Parametric Knowledge
LocationAgent通过分层推理和外部证据验证，有效提升图像地理定位的准确性和泛化能力。
扫码添加微信好友，提出您的宝贵建议 👇
💡 备注请填写：网站反馈
发表机构由 AI 从论文文本抽取，如发现错误，欢迎反馈修正建议。