🔥 Test-Time Gradient Guidance of Flow Policies in Reinforcement Learning

💡 The paper proposes a reinforcement learning algorithm called QGF that improves policies at test time by using a value gradient to guide a pre-trained flow policy. The problem addressed is that incorporating flow models into reinforcement learning pipelines for policy improvement can be difficult due to stability and scalability issues. The method involves pre-training a reference flow policy and a value function critic, then using the value gradient to guide the reference policy to generate higher-value actions at test time, without any additional policy learning. This approach avoids the instability of actor-critic training and sidesteps the need for specialized training objectives or backpropagating through denoising processes. The results show that QGF outperforms prior test-time reinforcement learning methods on single-task and goal-conditioned offline benchmarks with high-dimensional action spaces, and is competitive with state-of-the-art training-time algorithms while being much cheaper to run. Additionally, QGF exhibits favorable scaling with model size, offering a practical and effective alternative reinforcement learning algorithm with expressive policies. Overall, the paper contributes a new approach to reinforcement learning that improves policies at test time, avoiding training-time instability while maintaining competitive performance.

📅 Published on Jun 9

🔗 Links:
• GitHub: https://github.com/huggingface
• arXiv: https://arxiv.org/abs/2606.11087
• PDF: https://arxiv.org/pdf/2606.11087
• Project Page: https://q-guided-flow.github.io/

━━━━━━━━━━━━━━━━━━━━━━━━
📢 By: https://t.iss.one/PaperNexus

#ReinforcementLearningAlgorithms #FlowPolicyOptimization #TestTimePolicyImprovement #ValueGradientGuidance #QGFAlgorithm

🔥 Toward Generalist Autonomous Research via Hypothesis-Tree Refinement

💡 The paper introduces Arbor, a general framework for autonomous scientific research that enables an AI agent to run a repeated loop of exploration, experimentation, and abstraction over long horizons. The problem addressed is how to design an AI system that can conduct autonomous research across multiple domains, iteratively improving research outcomes without human supervision.

The method used is a combination of strategic coordination, isolated hypothesis testing, and a persistent knowledge tree, called Hypothesis Tree Refinement. The Arbor framework consists of a long-lived coordinator that manages global research strategy, short-lived executors that implement and test individual hypotheses, and a persistent tree that links hypotheses, artifacts, evidence, and distilled insights across time.

As results return, Arbor updates the tree, propagates reusable lessons, refines the search frontier, and admits verified improvements. This design turns autonomous research from a sequence of local attempts into a cumulative process in which strategy, execution, and evidence are carried across time.

The results show that Arbor achieves the best held-out result on all six real research tasks in model training, harness engineering, and data synthesis, attaining more than 2.5 times the average relative held-out gain of other methods under the same task interface and resource budget. Additionally, Arbor reaches 86.36 percent Any Medal with GPT-5.5 on MLE-Bench Lite, the strongest result in the comparison. Overall, the paper contributes a novel framework for autonomous research that demonstrates significant improvements in research outcomes across multiple domains.

📅 Published on Jun 10

🔗 Links:
• GitHub: https://github.com/huggingface
• arXiv: https://arxiv.org/abs/2606.11926
• PDF: https://arxiv.org/pdf/2606.11926
• Project Page: https://ruc-nlpir.github.io/Arbor/

━━━━━━━━━━━━━━━━━━━━━━━━
📢 By: https://t.iss.one/PaperNexus

#AutonomousScientificResearch #HypothesisTreeRefinement #ArtificialIntelligenceInScience #GeneralistAI #AutonomousResearchSystems

🔥 i1: A Simple and Fully Open Recipe for Strong Text-to-Image Models

💡 The paper presents a comprehensive study of text-to-image diffusion models, aiming to identify key design choices and training insights that lead to strong model performance. The problem addressed is the lack of fully open models that match the performance of state-of-the-art models, which hinders further research in the field. To tackle this, the authors conducted over 300 controlled experiments, totaling 700K TPU v6e hours, to investigate modeling and data design choices in text-to-image diffusion training and inference.

The method used involved a systematic investigation of various design decisions, such as dataset mixing and text encoder adapters, to identify simple yet effective approaches to training strong models. The authors found several empirical findings, including the use of equal weighting for mixing curated datasets and the benefits of larger text encoder adapters.

The results of the study led to the development of i1, a 3B-parameter text-to-image diffusion model trained using only publicly available datasets. The i1 model is competitive with leading models on five representative benchmarks and outperforms the best existing fully open model by 29.5 absolute percentage points on average. The authors provide the i1 checkpoints, training and inference code, and the data processing pipeline, making it a fully open model that can serve as a foundation for future research in text-to-image diffusion models.

Overall, the paper contributes to the field by providing a practical foundation for open research in text-to-image diffusion models, highlighting the importance of transparency and reproducibility in AI research. The release of the i1 model and its associated code and data processing pipeline enables the research community to build upon and improve the model, driving further progress in the field.

📅 Published on Jun 9

🔗 Links:
• GitHub: https://github.com/huggingface
• arXiv: https://arxiv.org/abs/2606.11289
• PDF: https://arxiv.org/pdf/2606.11289
• Project Page: https://zlab-princeton.github.io/i1/

🤖 Models citing this paper:
• https://huggingface.co/zlab-princeton/i1-3B

📊 Datasets citing this paper:
• https://huggingface.co/datasets/zlab-princeton/i1-captions
• https://huggingface.co/datasets/zlab-princeton/i1-gptedit-tfrecord

🚀 Spaces citing this paper:
• https://huggingface.co/spaces/multimodalart/i1-3B

━━━━━━━━━━━━━━━━━━━━━━━━
📢 By: https://t.iss.one/PaperNexus

#TextToImageModels #DiffusionModels #TextEncoderAdapters #ImageSynthesis #DeepLearningModels

🔥 DeNovoSWE: Scaling Long-Horizon Environments for Generating Entire Repositories from Scratch

💡 The paper introduces DeNovoSWE, a large scale dataset designed to train code agents to generate entire software repositories from documentation, which is a challenging long horizon software engineering task. The problem addressed in this paper is the scarcity of large scale datasets for whole repository generation, which hinders the training of code agents for such tasks. To address this issue, the authors propose a method to automatically construct a dataset through a sandboxed agentic workflow, which enables scalable curation without human annotation. The dataset is constructed using a divide and conquer approach and a critic repair philosophy, and a difficulty aware trajectory filtering strategy is used to balance data quality and diversity. The DeNovoSWE dataset comprises 4818 high quality instances, each requiring the generation of a complete repository from documentation. The results show that fine tuning a code agent on DeNovoSWE substantially improves its performance on long horizon software engineering tasks, with a significant increase in score on the BeyondSWE Doc2Repo benchmark from 5.8 percent to 47.2 percent. Overall, the paper contributes to the development of code agents that can generate entire software repositories from documentation, which has the potential to significantly improve software engineering productivity.

📅 Published on Jun 9

🔗 Links:
• GitHub: https://github.com/huggingface
• arXiv: https://arxiv.org/abs/2606.10728
• PDF: https://arxiv.org/pdf/2606.10728

📊 Datasets citing this paper:
• https://huggingface.co/datasets/AweAI-Team/DeNovoSWE
• https://huggingface.co/datasets/AweAI-Team/DeNovoSWE-Trajectory-Filtered
• https://huggingface.co/datasets/AweAI-Team/DeNovoSWE-Trajectory-Raw

━━━━━━━━━━━━━━━━━━━━━━━━
📢 By: https://t.iss.one/PaperNexus

#SoftwareRepositoryGeneration #CodeAgents #LongHorizonEnvironments #AutomatedDatasetConstruction #RepositoryGenerationFromDocumentation

🔥 TRACER: Trace-Based Adaptive Cost-Efficient Routing for LLM Classification

💡 The paper introduces TRACER, a system that uses production traces to train machine learning surrogates for large language model classification. The problem addressed is that large language models can be costly to deploy and may not always be necessary for every input. The authors propose a method where a lightweight surrogate model is trained on the production logs of the large language model, which can absorb a significant portion of future traffic at near-zero marginal inference cost.

The TRACER system trains the surrogate model on the production traces and governs its deployment through a parity gate, which activates the surrogate only when its agreement with the large language model exceeds a user-specified threshold. This approach provides interpretability into the handling boundaries of the surrogate model, describing which input regions it handles, where it plateaus, and why it defers.

The results show that TRACER achieves significant surrogate coverage on a 77-class intent benchmark, with coverage ranging from 83 to 100 percent depending on the quality target. On a 150-class benchmark, the surrogate fully replaces the teacher model. Additionally, the parity gate correctly refuses deployment on a natural language inference task when the embedding representation cannot support reliable separation. The TRACER system is available as open-source software, providing a cost-efficient and interpretable solution for large language model classification.

📅 Published on Apr 16

🔗 Links:
• GitHub: https://github.com/huggingface
• arXiv: https://arxiv.org/abs/2604.14531
• PDF: https://arxiv.org/pdf/2604.14531
• Project Page: https://www.tracer.deeprecall.io

━━━━━━━━━━━━━━━━━━━━━━━━
📢 By: https://t.iss.one/PaperNexus

#LLMClassification #TraceBasedRouting #AdaptiveCostEfficientRouting #LargeLanguageModelOptimization #MachineLearningSurrogates

🔥 Embodied-R1.5: Evolving Physical Intelligence via Embodied Foundation Models

💡 The paper introduces Embodied-R1.5, a unified embodied foundation model that integrates various embodied reasoning capabilities, such as cognition, task planning, correction, and pointing, into a single architecture. The goal is to achieve general physical intelligence. To train the model, the authors developed three automated data construction pipelines, resulting in a large-scale data system of over 15 billion tokens. They also designed a multi-task balanced reinforcement learning approach to alleviate conflicts between different tasks.

The model consists of a Planner-Grounder-Corrector framework, which enables it to autonomously execute and self-correct over long-horizon tasks. With only 8 billion parameters, Embodied-R1.5 achieves state-of-the-art performance on 16 out of 24 embodied vision-language benchmarks, surpassing leading models. The model can also be fine-tuned into a vision-language agent with a small amount of data, outperforming leading models across popular manipulation benchmark suites.

The authors conducted extensive zero-shot real-robot experiments, demonstrating the model's strong generalization to the physical world. The experiments validated the model's performance in instruction following, affordance grounding, articulated object manipulation, and long-horizon complex tasks. The paper's contributions include the introduction of the Embodied-R1.5 model, the development of a large-scale data system, and the creation of an evaluation framework tailored for embodied tasks. The model weights, datasets, training code, and evaluation framework are open-sourced to facilitate future research in embodied foundation models.

The problem addressed in the paper is the development of a unified embodied foundation model that can achieve general physical intelligence. The method used to address this problem is the integration of various embodied reasoning capabilities into a single architecture, along with the development of a large-scale data system and a multi-task balanced reinforcement learning approach. The results show that Embodied-R1.5 achieves state-of-the-art performance on various benchmarks and demonstrates strong generalization to the physical world. Overall, the paper contributes to the development of embodied foundation models and has the potential to facilitate future research in this area.

📅 Published on Jun 9

🔗 Links:
• GitHub: https://github.com/huggingface
• arXiv: https://arxiv.org/abs/2606.11324
• PDF: https://arxiv.org/pdf/2606.11324
• Project Page: https://embodied-r.github.io/

━━━━━━━━━━━━━━━━━━━━━━━━
📢 By: https://t.iss.one/PaperNexus

#EmbodiedIntelligence #PhysicalReasoning #FoundationModels #CognitiveArchitectures #ArtificialGeneralIntelligence

🔥 InterleaveThinker: Reinforcing Agentic Interleaved Generation

💡 The paper introduces InterleaveThinker, a multi-agent pipeline that enables existing image generators to perform interleaved generation, which involves generating a sequence of text and images. The current state-of-the-art image generators are limited in their ability to perform interleaved generation due to their architectures. InterleaveThinker addresses this limitation by using a planner agent to organize the input sequence and instruct the image generator, and a critic agent to evaluate the generator's outputs and refine the instructions. The pipeline is implemented using two modules, Interleave-Planner-SFT-80k and Interleave-Critic-SFT-112k, which are trained to perform a format cold-start, and Interleave-Critic-RL-13k, which is trained using reinforcement learning to correct instructions within a generation trajectory.

The method involves using a planner agent to plan the execution of the image generator at each step, and a critic agent to evaluate the generator's outputs and identify samples that deviate from the planned instructions. The critic agent then refines the instructions for regeneration. To optimize the entire generation trajectory, the authors propose using accuracy reward and step-wise reward, which allows single-step reinforcement learning to guide the entire trajectory.

The results show that InterleaveThinker improves the performance of various image generators on interleaved generation benchmarks, achieving performance comparable to state-of-the-art models such as Nano Banana and GPT-5. Additionally, InterleaveThinker significantly enhances the base model on reasoning-based benchmarks, such as 4-step FLUX.2-klein, where it achieves substantial gains on WISE and RISE. Overall, the paper demonstrates the effectiveness of InterleaveThinker in enabling existing image generators to perform interleaved generation and improving their performance on various benchmarks.

📅 Published on Jun 11

🔗 Links:
• GitHub: https://github.com/huggingface
• arXiv: https://arxiv.org/abs/2606.13679
• PDF: https://arxiv.org/pdf/2606.13679
• Project Page: https://zhengdian1.github.io/InterleaveThinker-proj/

🤖 Models citing this paper:
• https://huggingface.co/InterleaveThinker/InterleaveThinker-Planner-8B
• https://huggingface.co/InterleaveThinker/InterleaveThinker-Critic-8B
• https://huggingface.co/InterleaveThinker/Critic-SFT-8B

━━━━━━━━━━━━━━━━━━━━━━━━
📢 By: https://t.iss.one/PaperNexus

#InterleavedGeneration #AgenticInterleaving #MultiAgentPipelines #ImageTextGeneration #ReinforcedGeneration

🔥 Robust-U1: Can MLLMs Self-Recover Corrupted Visual Content for Robust Understanding?

💡 The paper proposes a novel framework called Robust-U1 to enhance the robustness of multimodal large language models against visual corruptions. The problem addressed is that existing models perform poorly when faced with real-world visual corruptions such as noise or blur. Current approaches to improve robustness have limitations, either lacking interpretability or being unable to restore lost pixel-level details.

The Robust-U1 framework is designed to equip models with explicit visual self-recovery capability, allowing them to recover corrupted visual content by themselves. The approach consists of three stages: supervised fine-tuning for initial reconstruction, reinforcement learning with dual rewards to align high visual quality, and multimodal reasoning that considers both the corrupted input and the recovered image.

The results show that Robust-U1 achieves state-of-the-art robustness on a real-world corruption benchmark and maintains superior performance under adversarial corruptions on general visual question answering benchmarks. The analysis confirms that high-quality visual recovery directly enhances reasoning performance, establishing self-recovery as a critical mechanism for robust visual understanding. Overall, the paper demonstrates that multimodal large language models can self-recover corrupted visual content, leading to improved robustness and performance in visual understanding tasks.

📅 Published on Jun 6

🔗 Links:
• GitHub: https://github.com/huggingface
• arXiv: https://arxiv.org/abs/2606.08063
• PDF: https://arxiv.org/pdf/2606.08063
• Project Page: https://huggingface.co/spaces/Jiaqi-hkust/Robust-U1

🤖 Models citing this paper:
• https://huggingface.co/Jiaqi-hkust/Robust-U1-SFT
• https://huggingface.co/Jiaqi-hkust/Robust-U1-RL
• https://huggingface.co/Jiaqi-hkust/Robust-U1

🚀 Spaces citing this paper:
• https://huggingface.co/spaces/Jiaqi-hkust/Robust-U1

━━━━━━━━━━━━━━━━━━━━━━━━
📢 By: https://t.iss.one/PaperNexus

#MultimodalLearning #VisualContentRecovery #RobustLanguageModels #SelfRecoveryMechanisms #CorruptionResistantAI

🔥 MiniMax Sparse Attention

💡 The paper introduces MiniMax Sparse Attention, a method for efficient processing of ultra-long contexts in large language models. The problem addressed is that the quadratic cost of softmax attention makes it difficult to jointly attend over hundreds of thousands to millions of tokens, which is necessary for applications such as agentic workflows, repository-scale code reasoning, and persistent memory.

To address this problem, the authors propose a blockwise sparse attention built upon Grouped Query Attention, called MiniMax Sparse Attention. This method consists of two branches: a lightweight Index Branch that scores key-value blocks and selects a Top-k subset for each group, and a Main Branch that performs exact block-sparse attention over only the selected blocks.

The method is designed to be simple and scalable, making it easy to deploy efficiently across a range of GPUs. The authors also co-design a GPU execution path that uses exp-free Top-k selection and KV-outer sparse attention to improve tensor-core utilization under block-granular access.

The results show that MiniMax Sparse Attention performs on par with Grouped Query Attention while reducing per-token attention compute by 28.4x at 1M context. When paired with the co-designed kernel, it achieves 14.2x prefill and 7.6x decoding wall-clock speedups on H800. The authors also release a production-grade natively multimodal model powered by MiniMax Sparse Attention, as well as the inference kernel, making it available for use by others.

Overall, the paper contributes a new method for efficient processing of ultra-long contexts in large language models, which has significant implications for applications that require joint attention over large numbers of tokens. The method is designed to be efficient, scalable, and easy to deploy, making it a valuable contribution to the field of natural language processing.

📅 Published on Jun 11

🔗 Links:
• GitHub: https://github.com/huggingface
• arXiv: https://arxiv.org/abs/2606.13392
• PDF: https://arxiv.org/pdf/2606.13392

🤖 Models citing this paper:
• https://huggingface.co/MiniMaxAI/MiniMax-M3
• https://huggingface.co/MiniMaxAI/MiniMax-M3-MXFP8
• https://huggingface.co/sparkarena/Minimax-M3-v0-NVFP4

🚀 Spaces citing this paper:
• https://huggingface.co/spaces/saivivek6/updated_mongodb_p
• https://huggingface.co/spaces/akhaliq/MiniMax-M3

━━━━━━━━━━━━━━━━━━━━━━━━
📢 By: https://t.iss.one/PaperNexus

#MiniMaxSparseAttention #EfficientLanguageModeling #SparseAttentionMechanisms #UltraLongContextProcessing #BlockwiseAttentionMethods

🔥 SpatialClaw: Rethinking Action Interface for Agentic Spatial Reasoning

💡 The paper introduces SpatialClaw, a training-free framework that enables flexible and stateful spatial reasoning in vision-language models. The problem addressed is the limitation of current spatial agents in performing open-ended spatial reasoning tasks, which is due to the design of the action interface that invokes specialist perception modules. Existing spatial agents use either single-pass code execution or a structured tool-call interface, both of which offer limited flexibility for complex 3D/4D spatial reasoning.

The proposed SpatialClaw framework uses code as the action interface, allowing a vision-language model-backed agent to write executable code conditioned on prior outputs. This approach enables the agent to flexibly compose and manipulate perception results and adapt its analysis to intermediate text and visual observations. SpatialClaw maintains a stateful Python kernel pre-loaded with input frames and a suite of perception and geometry primitives.

The results show that SpatialClaw achieves superior performance across diverse 3D/4D spatial reasoning tasks, with an average accuracy of 59.9% across 20 benchmarks. This represents a significant improvement of 11.2 points over the recent spatial agent, with consistent gains across six vision-language model backbones from two model families, without any benchmark- or model-specific adaptation. The paper's contribution is the introduction of a flexible and effective framework for spatial reasoning that can be applied to a wide range of tasks without requiring training or adaptation.

📅 Published on Jun 11

🔗 Links:
• GitHub: https://github.com/huggingface
• arXiv: https://arxiv.org/abs/2606.13673
• PDF: https://arxiv.org/pdf/2606.13673
• Project Page: https://spatialclaw.github.io/

━━━━━━━━━━━━━━━━━━━━━━━━
📢 By: https://t.iss.one/PaperNexus

#SpatialReasoning #VisionLanguageModels #AgenticInterfaces #SpatialArtificialIntelligence #CodeBasedActionInterfaces

🔥 EurekAgent: Agent Environment Engineering is All You Need For Autonomous Scientific Discovery

💡 The paper introduces EurekAgent, a system designed to enhance autonomous scientific discovery through environment engineering. The authors argue that as model capabilities improve, the main bottleneck for autonomous scientific discovery shifts from designing agent workflows to designing agent environments. Environment engineering involves building environments that promote productive behaviors such as exploration, collaboration, and artifact management, while suppressing harmful behaviors like reward hacking and high-friction human oversight.

The EurekAgent system engineers the environment along four dimensions: permissions engineering, artifact engineering, budget engineering, and human-in-the-loop engineering. Permissions engineering allows for bounded agent execution and isolated evaluation, while artifact engineering enables filesystem and Git-based collaboration. Budget engineering enables budget-aware exploration, and human-in-the-loop engineering facilitates easy human supervision and intervention.

The authors demonstrate the effectiveness of EurekAgent by achieving state-of-the-art results across multiple domains, including mathematics, kernel engineering, and machine learning tasks. Notably, EurekAgent discovered new state-of-the-art 26-circle packing results with less than 11 dollars in total API cost. The system's low computational costs and impressive results highlight the potential of environment engineering for autonomous scientific discovery.

The paper's main contribution is the introduction of environment engineering as a core research direction for developing reliable autonomous research agents. By open-sourcing their code and results, the authors invite the research community to explore and build upon their work, with the goal of advancing autonomous scientific discovery. Overall, the paper presents a significant step forward in the development of autonomous scientific discovery systems, and highlights the importance of environment engineering in achieving reliable and efficient autonomous research.

📅 Published on Jun 11

🔗 Links:
• GitHub: https://github.com/huggingface
• arXiv: https://arxiv.org/abs/2606.13662
• PDF: https://arxiv.org/pdf/2606.13662

━━━━━━━━━━━━━━━━━━━━━━━━
📢 By: https://t.iss.one/PaperNexus

#AutonomousScientificDiscovery #AgentEnvironmentEngineering #EnvironmentEngineering #ArtificialIntelligenceForScience #AutonomousResearchSystems