PyTorchKR Papers

최신 논문

총 1,904편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2504.10733

Cross-Problem Parameter Transfer in Quantum Approximate Optimization Algorithm: A Machine Learning Approach

Quantum Approximate Optimization Algorithm (QAOA) is one of the most promising candidates to achieve the quantum advantage in solving combinatorial optimization problems. The process of finding a good set of variational parameters in the QAOA circuit has proven to be challenging due to multiple factors, such as barren plateaus. As a result, there is growing interest in exploiting parameter transferability, where parameter sets optimized for one problem instance are transferred to another that could be more complex either to estimate the solution or to serve as a warm start for further optimization. But can we transfer parameters from one class of problems to another? Leveraging parameter sets learned from a well-studied class of problems could help navigate the less studied one, reducing optimization overhead and mitigating performance pitfalls. In this paper, we study whether pretrained QAOA parameters of MaxCut can be used as is or to warm start the Maximum Independent Set (MIS) circuits. Specifically, we design machine learning models to find good donor candidates optimized on MaxCut and apply their parameters to MIS acceptors. Our experimental results show that such parameter transfer can significantly reduce the number of optimization iterations required while achieving comparable approximation ratios.

arXiv:2504.09843

ST-Booster: An Iterative SpatioTemporal Perception Booster for Vision-and-Language Navigation in Continuous Environments

Vision-and-Language Navigation in Continuous Environments (VLN-CE) requires agents to navigate unknown, continuous spaces based on natural language instructions. Compared to discrete settings, VLN-CE poses two core perception challenges. First, the absence of predefined observation points leads to heterogeneous visual memories and weakened global spatial correlations. Second, cumulative reconstruction errors in three-dimensional scenes introduce structural noise, impairing local feature perception. To address these challenges, this paper proposes ST-Booster, an iterative spatiotemporal booster that enhances navigation performance through multi-granularity perception and instruction-aware reasoning. ST-Booster consists of three key modules -- Hierarchical SpatioTemporal Encoding (HSTE), Multi-Granularity Aligned Fusion (MGAF), and ValueGuided Waypoint Generation (VGWG). HSTE encodes long-term global memory using topological graphs and captures shortterm local details via grid maps. MGAF aligns these dualmap representations with instructions through geometry-aware knowledge fusion. The resulting representations are iteratively refined through pretraining tasks. During reasoning, VGWG generates Guided Attention Heatmaps (GAHs) to explicitly model environment-instruction relevance and optimize waypoint selection. Extensive comparative experiments and performance analyses are conducted, demonstrating that ST-Booster outperforms existing state-of-the-art methods, particularly in complex, disturbance-prone environments.

arXiv:2504.09532

Embodied Chain of Action Reasoning with Multi-Modal Foundation Model for Humanoid Loco-manipulation

Enabling humanoid robots to autonomously perform loco-manipulation tasks in complex, unstructured environments poses significant challenges. This entails equipping robots with the capability to plan actions over extended horizons while leveraging multi-modality to bridge gaps between high-level planning and actual task execution. Recent advancements in multi-modal foundation models have showcased substantial potential in enhancing planning and reasoning abilities, particularly in the comprehension and processing of semantic information for robotic control tasks. In this paper, we introduce a novel framework based on foundation models that applies the embodied chain of action reasoning methodology to autonomously plan actions from textual instructions for humanoid loco-manipulation. Our method integrates humanoid-specific chain of thought methodology, including detailed affordance and body movement analysis, which provides a breakdown of the task into a sequence of locomotion and manipulation actions. Moreover, we incorporate spatial reasoning based on the observation and target object properties to effectively navigate where target position may be unseen or occluded. Through rigorous experimental setups on object rearrangement, manipulations and loco-manipulation tasks on a real-world environment, we evaluate our method's efficacy on the decoupled upper and lower body control and demonstrate the effectiveness of the chain of robotic action reasoning strategies in comprehending human instructions.

arXiv:2504.10620

SPreV

SPREV, short for hyperSphere Reduced to two-dimensional Regular Polygon for Visualisation, is a novel dimensionality reduction technique developed to address the challenges of reducing dimensions and visualizing labeled datasets that exhibit a unique combination of three characteristics: small class size, high dimensionality, and low sample size. SPREV is designed not only to uncover but also to visually represent hidden patterns within such datasets. Its distinctive integration of geometric principles, adapted for discrete computational environments, makes it an indispensable tool in the modern data science toolkit, enabling users to identify trends, extract insights, and navigate complex data efficiently and effectively.

arXiv:2504.09775

Understanding and Optimizing Multi-Stage AI Inference Pipelines

The rapid evolution of Large Language Models (LLMs) has driven the need for increasingly sophisticated inference pipelines and hardware platforms. Modern LLM serving extends beyond traditional prefill-decode workflows, incorporating multi-stage processes such as Retrieval Augmented Generation (RAG), key-value (KV) cache retrieval, dynamic model routing, and multi step reasoning. These stages exhibit diverse computational demands, requiring distributed systems that integrate GPUs, ASICs, CPUs, and memory-centric architectures. However, existing simulators lack the fidelity to model these heterogeneous, multi-engine workflows, limiting their ability to inform architectural decisions. To address this gap, we introduce HERMES, a Heterogeneous Multi-stage LLM inference Execution Simulator. HERMES models diverse request stages; including RAG, KV retrieval, reasoning, prefill, and decode across complex hardware hierarchies. HERMES supports heterogeneous clients executing multiple models concurrently unlike prior frameworks while incorporating advanced batching strategies and multi-level memory hierarchies. By integrating real hardware traces with analytical modeling, HERMES captures critical trade-offs such as memory bandwidth contention, inter-cluster communication latency, and batching efficiency in hybrid CPU-accelerator deployments. Through case studies, we explore the impact of reasoning stages on end-to-end latency, optimal batching strategies for hybrid pipelines, and the architectural implications of remote KV cache retrieval. HERMES empowers system designers to navigate the evolving landscape of LLM inference, providing actionable insights into optimizing hardware-software co-design for next-generation AI workloads.

arXiv:2504.08581

FMLGS: Fast Multilevel Language Embedded Gaussians for Part-level Interactive Agents

The semantically interactive radiance field has long been a promising backbone for 3D real-world applications, such as embodied AI to achieve scene understanding and manipulation. However, multi-granularity interaction remains a challenging task due to the ambiguity of language and degraded quality when it comes to queries upon object components. In this work, we present FMLGS, an approach that supports part-level open-vocabulary query within 3D Gaussian Splatting (3DGS). We propose an efficient pipeline for building and querying consistent object- and part-level semantics based on Segment Anything Model 2 (SAM2). We designed a semantic deviation strategy to solve the problem of language ambiguity among object parts, which interpolates the semantic features of fine-grained targets for enriched information. Once trained, we can query both objects and their describable parts using natural language. Comparisons with other state-of-the-art methods prove that our method can not only better locate specified part-level targets, but also achieve first-place performance concerning both speed and accuracy, where FMLGS is 98 x faster than LERF, 4 x faster than LangSplat and 2.5 x faster than LEGaussians. Meanwhile, we further integrate FMLGS as a virtual agent that can interactively navigate through 3D scenes, locate targets, and respond to user demands through a chat interface, which demonstrates the potential of our work to be further expanded and applied in the future.

arXiv:2504.08077

Lemons and Learning: Adult and Juvenile Free-Ranging Dogs Navigate Aversive Foraging Challenges Differently

Understanding how animals make foraging decisions in challenging or unpleasant contexts sheds light on the processes that underlie cognitive development and the evolution of adaptive foraging techniques in complex ecological settings. In order to investigate how age, gender, and environmental cues impact decision-making and scavenging behaviour, this study explores how juvenile and adult free-ranging dogs (FRDs) respond to aversive stimuli, specifically different concentrations of lemon juice. We conducted a three-bowl choice experiment using lemon juice at concentrations of 25%, 33.3%, and 50% with 73 juvenile FRDs. The juvenile FRD data were compared with adult behavioural data from a prior experiment. The findings indicated that juvenile dogs' foraging behaviour was consistent under provided circumstances, suggesting that their capacity to adapt to aversive acidic stimuli was limited, possibly due to their inexperience and developing cognitive abilities. Adult dogs, on the other hand, exhibited selective foraging by preferring lower acidity and demonstrating more deliberate foraging strategies. Age, sex, and acidity all had significant impact on eating behaviour, with males displaying a higher tolerance to unpleasant stimuli. Markov chain analysis showed repetitive decision-making patterns, especially in adults, indicating continuous sensory evaluation before food consumption, while strategy-making behaviour increased the likelihood of food consumption in juveniles, suggesting juveniles are possibly beginning to build adaptive skills. These findings demonstrate the dynamic interaction of developmental stage, sensory processing, and individual diversity, in developing of foraging strategies of FRDs. This offers new perspectives to understand the cognitive basis of scavenging behaviour in aversive ecological conditions like urban and human-dominated landscapes.

arXiv:2504.04365

AutoPDL: Automatic Prompt Optimization for LLM Agents

The performance of large language models (LLMs) depends on how they are prompted, with choices spanning both the high-level prompting pattern (e.g., Zero-Shot, CoT, ReAct, ReWOO) and the specific prompt content (instructions and few-shot demonstrations). Manually tuning this combination is tedious, error-prone, and specific to a given LLM and task. Therefore, this paper proposes AutoPDL, an automated approach to discovering good LLM agent configurations. Our approach frames this as a structured AutoML problem over a combinatorial space of agentic and non-agentic prompting patterns and demonstrations, using successive halving to efficiently navigate this space. We introduce a library implementing common prompting patterns using the PDL prompt programming language. AutoPDL solutions are human-readable, editable, and executable PDL programs that use this library. This approach also enables source-to-source optimization, allowing human-in-the-loop refinement and reuse. Evaluations across three tasks and seven LLMs (ranging from 3B to 70B parameters) show consistent accuracy gains ($9.06\pm15.3$ percentage points), up to 68.9pp, and reveal that selected prompting strategies vary across models and tasks.

arXiv:2504.05333

When is using AI the rational choice? The importance of counterfactuals in AI deployment decisions

Decisions to deploy AI capabilities are often driven by counterfactuals - a comparison of decisions made using AI to decisions that would have been made if the AI were not used. Counterfactual misses, which are poor decisions that are attributable to using AI, may have disproportionate disutility to AI deployment decision makers. Counterfactual hits, which are good decisions attributable to AI usage, may provide little benefit beyond the benefit of better decisions. This paper explores how to include counterfactual outcomes into usage decision expected utility assessments. Several properties emerge when counterfactuals are explicitly included. First, there are many contexts where the expected utility of AI usage is positive for intended beneficiaries and strongly negative for stakeholders and deployment decision makers. Second, high levels of complementarity, where differing AI and user assessments are merged beneficially, often leads to substantial disutility for stakeholders. Third, apparently small changes in how users interact with an AI capability can substantially impact stakeholder utility. Fourth, cognitive biases such as expert overconfidence and hindsight bias exacerbate the perceived frequency of costly counterfactual misses. The expected utility assessment approach presented here is intended to help AI developers and deployment decision makers to navigate the subtle but substantial impact of counterfactuals so as to better ensure that beneficial AI capabilities are used.

arXiv:2504.04517공식 코드 · Papers with Code

Enhance Then Search: An Augmentation-Search Strategy with Foundation Models for Cross-Domain Few-Shot Object Detection

Foundation models pretrained on extensive datasets, such as GroundingDINO and LAE-DINO, have performed remarkably in the cross-domain few-shot object detection (CD-FSOD) task. Through rigorous few-shot training, we found that the integration of image-based data augmentation techniques and grid-based sub-domain search strategy significantly enhances the performance of these foundation models. Building upon GroundingDINO, we employed several widely used image augmentation methods and established optimization objectives to effectively navigate the expansive domain space in search of optimal sub-domains. This approach facilitates efficient few-shot object detection and introduces an approach to solving the CD-FSOD problem by efficiently searching for the optimal parameter configuration from the foundation model. Our findings substantially advance the practical deployment of vision-language models in data-scarce environments, offering critical insights into optimizing their cross-domain generalization capabilities without labor-intensive retraining. Code is available at https://github.com/jaychempan/ETS.

arXiv:2504.03160공식 코드 · Papers with Code

DeepResearcher: Scaling Deep Research via Reinforcement Learning in Real-world Environments

Large Language Models (LLMs) equipped with web search capabilities have demonstrated impressive potential for deep research tasks. However, current approaches predominantly rely on either manually engineered prompts (prompt engineering-based) with brittle performance or reinforcement learning within controlled Retrieval-Augmented Generation (RAG) environments (RAG-based) that fail to capture the complexities of real-world interaction. In this paper, we introduce DeepResearcher, the first comprehensive framework for end-to-end training of LLM-based deep research agents through scaling reinforcement learning (RL) in real-world environments with authentic web search interactions. Unlike RAG-based approaches that assume all necessary information exists within a fixed corpus, our method trains agents to navigate the noisy, unstructured, and dynamic nature of the open web. We implement a specialized multi-agent architecture where browsing agents extract relevant information from various webpage structures and overcoming significant technical challenges. Extensive experiments on open-domain research tasks demonstrate that DeepResearcher achieves substantial improvements of up to 28.9 points over prompt engineering-based baselines and up to 7.2 points over RAG-based RL agents. Our qualitative analysis reveals emergent cognitive behaviors from end-to-end RL training, including the ability to formulate plans, cross-validate information from multiple sources, engage in self-reflection to redirect research, and maintain honesty when unable to find definitive answers. Our results highlight that end-to-end training in real-world web environments is not merely an implementation detail but a fundamental requirement for developing robust research capabilities aligned with real-world applications. We release DeepResearcher at https://github.com/GAIR-NLP/DeepResearcher.

arXiv:2504.03497

Hybrid Real- and Complex-valued Neural Network Architecture

We propose a \emph{hybrid} real- and complex-valued \emph{neural network} (HNN) architecture, designed to combine the computational efficiency of real-valued processing with the ability to effectively handle complex-valued data. We illustrate the limitations of using real-valued neural networks (RVNNs) for inherently complex-valued problems by showing how it learnt to perform complex-valued convolution, but with notable inefficiencies stemming from its real-valued constraints. To create the HNN, we propose to use building blocks containing both real- and complex-valued paths, where information between domains is exchanged through domain conversion functions. We also introduce novel complex-valued activation functions, with higher generalisation and parameterisation efficiency. HNN-specific architecture search techniques are described to navigate the larger solution space. Experiments with the AudioMNIST dataset demonstrate that the HNN reduces cross-entropy loss and consumes less parameters compared to an RVNN for all considered cases. Such results highlight the potential for the use of partially complex-valued processing in neural networks and applications for HNNs in many signal processing domains.

arXiv:2504.03561공식 코드 · Papers with Code

SynWorld: Virtual Scenario Synthesis for Agentic Action Knowledge Refinement

In the interaction between agents and their environments, agents expand their capabilities by planning and executing actions. However, LLM-based agents face substantial challenges when deployed in novel environments or required to navigate unconventional action spaces. To empower agents to autonomously explore environments, optimize workflows, and enhance their understanding of actions, we propose SynWorld, a framework that allows agents to synthesize possible scenarios with multi-step action invocation within the action space and perform Monte Carlo Tree Search (MCTS) exploration to effectively refine their action knowledge in the current environment. Our experiments demonstrate that SynWorld is an effective and general approach to learning action knowledge in new environments. Code is available at https://github.com/zjunlp/SynWorld.

arXiv:2404.15209

Data-Driven Knowledge Transfer in Batch $Q^*$ Learning

In data-driven decision-making in marketing, healthcare, and education, it is desirable to utilize a large amount of data from existing ventures to navigate high-dimensional feature spaces and address data scarcity in new ventures. We explore knowledge transfer in dynamic decision-making by concentrating on batch stationary environments and formally defining task discrepancies through the lens of Markov decision processes (MDPs). We propose a framework of Transferred Fitted $Q$-Iteration algorithm with general function approximation, enabling the direct estimation of the optimal action-state function $Q^*$ using both target and source data. We establish the relationship between statistical performance and MDP task discrepancy under sieve approximation, shedding light on the impact of source and target sample sizes and task discrepancy on the effectiveness of knowledge transfer. We show that the final learning error of the $Q^*$ function is significantly improved from the single task rate both theoretically and empirically.

arXiv:2504.01408공식 코드 · Papers with Code

From Shadows to Safety: Occlusion Tracking and Risk Mitigation for Urban Autonomous Driving

Autonomous vehicles (AVs) must navigate dynamic urban environments where occlusions and perception limitations introduce significant uncertainties. This research builds upon and extends existing approaches in risk-aware motion planning and occlusion tracking to address these challenges. While prior studies have developed individual methods for occlusion tracking and risk assessment, a comprehensive method integrating these techniques has not been fully explored. We, therefore, enhance a phantom agent-centric model by incorporating sequential reasoning to track occluded areas and predict potential hazards. Our model enables realistic scenario representation and context-aware risk evaluation by modeling diverse phantom agents, each with distinct behavior profiles. Simulations demonstrate that the proposed approach improves situational awareness and balances proactive safety with efficient traffic flow. While these results underline the potential of our method, validation in real-world scenarios is necessary to confirm its feasibility and generalizability. By utilizing and advancing established methodologies, this work contributes to safer and more reliable AV planning in complex urban environments. To support further research, our method is available as open-source software at: https://github.com/TUM-AVS/OcclusionAwareMotionPlanning

arXiv:2504.01089

HomeEmergency -- Using Audio to Find and Respond to Emergencies in the Home

In the United States alone accidental home deaths exceed 128,000 per year. Our work aims to enable home robots who respond to emergency scenarios in the home, preventing injuries and deaths. We introduce a new dataset of household emergencies based in the ThreeDWorld simulator. Each scenario in our dataset begins with an instantaneous or periodic sound which may or may not be an emergency. The agent must navigate the multi-room home scene using prior observations, alongside audio signals and images from the simulator, to determine if there is an emergency or not. In addition to our new dataset, we present a modular approach for localizing and identifying potential home emergencies. Underpinning our approach is a novel probabilistic dynamic scene graph (P-DSG), where our key insight is that graph nodes corresponding to agents can be represented with a probabilistic edge. This edge, when refined using Bayesian inference, enables efficient and effective localization of agents in the scene. We also utilize multi-modal vision-language models (VLMs) as a component in our approach, determining object traits (e.g. flammability) and identifying emergencies. We present a demonstration of our method completing a real-world version of our task on a consumer robot, showing the transferability of both our task and our method. Our dataset will be released to the public upon this papers publication.

arXiv:2503.24140

Reinforcement Learning for Safe Autonomous Two Device Navigation of Cerebral Vessels in Mechanical Thrombectomy

Purpose: Autonomous systems in mechanical thrombectomy (MT) hold promise for reducing procedure times, minimizing radiation exposure, and enhancing patient safety. However, current reinforcement learning (RL) methods only reach the carotid arteries, are not generalizable to other patient vasculatures, and do not consider safety. We propose a safe dual-device RL algorithm that can navigate beyond the carotid arteries to cerebral vessels. Methods: We used the Simulation Open Framework Architecture to represent the intricacies of cerebral vessels, and a modified Soft Actor-Critic RL algorithm to learn, for the first time, the navigation of micro-catheters and micro-guidewires. We incorporate patient safety metrics into our reward function by integrating guidewire tip forces. Inverse RL is used with demonstrator data on 12 patient-specific vascular cases. Results: Our simulation demonstrates successful autonomous navigation within unseen cerebral vessels, achieving a 96% success rate, 7.0s procedure time, and 0.24 N mean forces, well below the proposed 1.5 N vessel rupture threshold. Conclusion: To the best of our knowledge, our proposed autonomous system for MT two-device navigation reaches cerebral vessels, considers safety, and is generalizable to unseen patient-specific cases for the first time. We envisage future work will extend the validation to vasculatures of different complexity and on in vitro models. While our contributions pave the way towards deploying agents in clinical settings, safety and trustworthiness will be crucial elements to consider when proposing new methodology.

arXiv:2503.23105

Open-Vocabulary Semantic Segmentation with Uncertainty Alignment for Robotic Scene Understanding in Indoor Building Environments

The global rise in the number of people with physical disabilities, in part due to improvements in post-trauma survivorship and longevity, has amplified the demand for advanced assistive technologies to improve mobility and independence. Autonomous assistive robots, such as smart wheelchairs, require robust capabilities in spatial segmentation and semantic recognition to navigate complex built environments effectively. Place segmentation involves delineating spatial regions like rooms or functional areas, while semantic recognition assigns semantic labels to these regions, enabling accurate localization to user-specific needs. Existing approaches often utilize deep learning; however, these close-vocabulary detection systems struggle to interpret intuitive and casual human instructions. Additionally, most existing methods ignore the uncertainty of the scene recognition problem, leading to low success rates, particularly in ambiguous and complex environments. To address these challenges, we propose an open-vocabulary scene semantic segmentation and detection pipeline leveraging Vision Language Models (VLMs) and Large Language Models (LLMs). Our approach follows a 'Segment Detect Select' framework for open-vocabulary scene classification, enabling adaptive and intuitive navigation for assistive robots in built environments.

arXiv:2503.24282

Style Quantization for Data-Efficient GAN Training

Under limited data setting, GANs often struggle to navigate and effectively exploit the input latent space. Consequently, images generated from adjacent variables in a sparse input latent space may exhibit significant discrepancies in realism, leading to suboptimal consistency regularization (CR) outcomes. To address this, we propose \textit{SQ-GAN}, a novel approach that enhances CR by introducing a style space quantization scheme. This method transforms the sparse, continuous input latent space into a compact, structured discrete proxy space, allowing each element to correspond to a specific real data point, thereby improving CR performance. Instead of direct quantization, we first map the input latent variables into a less entangled ``style'' space and apply quantization using a learnable codebook. This enables each quantized code to control distinct factors of variation. Additionally, we optimize the optimal transport distance to align the codebook codes with features extracted from the training data by a foundation model, embedding external knowledge into the codebook and establishing a semantically rich vocabulary that properly describes the training dataset. Extensive experiments demonstrate significant improvements in both discriminator robustness and generation quality with our method.

arXiv:2503.23271

Learning Coordinated Bimanual Manipulation Policies using State Diffusion and Inverse Dynamics Models

When performing tasks like laundry, humans naturally coordinate both hands to manipulate objects and anticipate how their actions will change the state of the clothes. However, achieving such coordination in robotics remains challenging due to the need to model object movement, predict future states, and generate precise bimanual actions. In this work, we address these challenges by infusing the predictive nature of human manipulation strategies into robot imitation learning. Specifically, we disentangle task-related state transitions from agent-specific inverse dynamics modeling to enable effective bimanual coordination. Using a demonstration dataset, we train a diffusion model to predict future states given historical observations, envisioning how the scene evolves. Then, we use an inverse dynamics model to compute robot actions that achieve the predicted states. Our key insight is that modeling object movement can help learning policies for bimanual coordination manipulation tasks. Evaluating our framework across diverse simulation and real-world manipulation setups, including multimodal goal configurations, bimanual manipulation, deformable objects, and multi-object setups, we find that it consistently outperforms state-of-the-art state-to-action mapping policies. Our method demonstrates a remarkable capacity to navigate multimodal goal configurations and action distributions, maintain stability across different control modes, and synthesize a broader range of behaviors than those present in the demonstration dataset.