FuguReport

2026-09-04 Method / Domain Generalization / Risk-aware adversarial tuning and generation

PAPT++: Risk-Aware Adversarial Tuning and Generation for Single Domain Generalization

PAPT++ addresses single domain generalization by using a pretrained text-to-image diffusion model to generate training samples that are not only diverse but also difficult for the current classifier.

2026-09-04 Evaluation / Reasoning Assessment / Mathematical reasoning knowledge evaluation

Do LLMs Exhibit Coherent Knowledge Structures in Mathematical Reasoning? A Perspective from Knowledge Space Theory

This paper proposes a Knowledge Space Theory (KST)-grounded framework to evaluate whether large language models exhibit coherent knowledge structures in mathematical reasoning rather than only high answer accuracy.

2026-09-04 Method / Model Explanation / Causal dependency graph construction

DCFA: Dual-view Causal-inspired Attribution for Failure Reasoning in LLM-based Multi-agent Systems

This paper addresses failure attribution in large language model (LLM)-based multi-agent systems, aiming to identify the decisive error whose correction reverses or maximally reduces system failure.

2026-09-04 Method / Agent Environments / Scalable environment pipeline

CUA-Universe: A Scalable and Dynamic Environment for Hybrid GUI+CLI Agents

CUA-Universe is a pipeline for turning real desktop software into hybrid environments where agents can act through both graphical interfaces and command-line tools over shared application state.

2026-09-04 Method / World Model / Physics-informed world modeling

TourPhysics: Bringing Physics to World Models for Exploration and Manipulation from a Single Image

TourPhysics is an online world-model framework that initializes a controllable scene hypothesis from a single image and a declarative physical configuration.

2026-09-03 Method / Test-Time Adaptation / Adaptation via human-AI interaction

Efficient Test-Time Adaptation through Human-AI Interaction

This paper introduces TAHI (Test-time Adaptation through Human-agent Interaction), a framework enabling interactive AI agents to adapt to individual human expertise on open-ended tasks.

2026-09-03 Method / Attention / Random token preservation in attention heads

Random Attention: Rethinking KV Cache Eviction for Efficient Reasoning

This paper studies KV-cache eviction for long-chain reasoning in large language models and questions the standard assumption that cached tokens must be ranked by an importance score.

2026-09-03 Method / Diffusion Models / Training diffusion transformer from scratch

LLaDA-Image: Building Strong Image Generators with Fully Open Training Recipes

LLaDA-Image is a unified image generation framework centered on a 6B Diffusion Transformer trained from scratch and conditioned by a frozen dLLM-based vision-language module.

2026-09-03 Method / Embedding / Distillation for compositional reasoning

CORE: Improving Compositional Reasoning in MLLM Embedding via Reranker Distillation

This paper studies a gap between cross-attentive rerankers and dual-encoder embeddings in multimodal retrieval: rerankers can often make fine-grained compositional judgments that are not preserved in embedding space.

2026-09-03 Method / Pose Estimation / Multi-relative pose query learning

Scal3R: Learning Efficient Multi-Relative Pose Query for Scalable Online 3D Reconstruction

Scal3R addresses the failure of online 3D reconstruction models on long video sequences by replacing absolute pose regression to a fixed first frame with multi-reference relative pose queries.

2026-09-02 Method / Video Modeling / Interactive video world model construction

SolarWM: Open Data and Scalable Training for Long-Horizon Video World Models

SolarWM is introduced as an open foundation for interactive video world models that combines an open data engine with a unified training stack for heterogeneous video backbones.

2026-09-02 Method / Model Fine-Tuning / Post-training with SFT and RL

Post-Training Language Models for Gold-Medal Performance in Coding Competitions

This paper presents an end-to-end post-training and inference pipeline for competitive programming that integrates large-scale problem curation, synthetic reasoning traces, supervised fine-tuning (SFT), reinforcement learning (RL), and an iterative test-time compute procedure called GenCorrect.

2026-09-02 Method / Attention / Controlling attention in decoding

Language Models Can Control Their Own Attention

This paper introduces Declarative Attention (DA), a prompting and inference protocol in which a language model explicitly declares where it needs to attend during decoding.

2026-09-02 Method / Reinforcement Learning / Tail-likelihood based learning

Tail-Likelihood Reinforcement Learning

This paper proposes Tail-Likelihood Reinforcement Learning (TailRL), an RL objective for continuous rewards that maximizes the expected log-probability of exceeding a uniformly sampled reward threshold.

2026-09-02 Method / In-Context Learning / Task-specific reasoning operation learning

SALA: Semantic-Aware Logical Alignment for Complex Reasoning in In-Context Learning

This paper studies demonstration selection for complex reasoning in in-context learning, arguing that surface-level similarity and rigid predefined reasoning-step matching do not adequately capture problem-solving logic.

2026-09-01 Method / Model Training / Training personalized simulators

StudentSim: Training LLM-based Student Simulators

StudentSim is a two-stage training framework for building individualized student simulators from sparse per-student data.

2026-09-01 Method / Reinforcement Learning / Off-policy and on-policy reward supervision

Scaling Near-Optimal SFT-RL Annotation Budget Allocation from Small to Large LLMs

This paper studies how to split a fixed post-training annotation budget between supervised fine-tuning (SFT) and a subsequent RL stage for large language models.

2026-09-01 Method / Multimodal Models / Unified understanding-generation framework

Uncovering Understanding-Generation Synergy in Native Unified Multimodal Models: From Representation, Task to System

This paper studies whether unified multimodal models actually gain from jointly learning visual understanding and image generation, rather than merely exposing both functions through one interface.

2026-09-01 Method / Knowledge Distillation / Mid-training distillation with confident tokens

Knowledge Distillation During Mid-Training Favors Reasoning over Factual Recall

This paper studies how logit-based knowledge distillation behaves during mid-training, the continued self-supervised stage between pre-training and post-training.

2026-09-01 Method / Sparse Models / Input-adaptive sparse prefilling technique

CRISP: Cliff-awaRe Input-adaptive Sparse Prefilling with Structural-Mass-Motivated Routing

CRISP is a training-free method for long-context sparse prefilling that targets two limitations identified in prior dynamic sparse attention systems.

2026-08-31 Evaluation / Benchmarking / Multi-turn diagram generation benchmark

PaperBanana-Interact: Scientific Diagram Refinement with Multi-Turn Human Feedback

This paper studies multi-turn refinement of scientific diagrams, arguing that single-pass generation often fails to satisfy an author's evolving visual preferences.

2026-08-31 Evaluation / Benchmarking / Partially observable MNIST benchmark

MNIST-PRO: MNIST is Back as a Partially Observable World for AI Agents

MNIST-PRO reformulates MNIST digit recognition as a partially observable sequential search task in which an agent must move a glimpse window, accumulate evidence, and decide when to predict.

2026-08-31 Method / Research Automation / Framework for converting papers to training environments

PaperGym: Rubric-Centered Evolution for Research-Plan Generation

PaperGym is a framework that converts scientific papers into training environments for research-plan generation.

2026-08-31 Method / Quantization / Consistent low-bit fine-tuning

Fine-Tuning Low-Bit Models with Gradient in Quantized Code Space

This paper studies strict fully low-bit fine-tuning, where the adapted model must remain in the same quantized format at deployment, so optimization is performed over quantization codes and scales rather than a continuous weight space.

2026-08-30 Method / Multi-task Learning / Joint training on speech and music timestamps

TEMPO: Temporally-grounded Multi-task Post-training for Large Audio-Language Models

The paper addresses a central limitation of large audio-language models: they typically describe an entire clip but do not assign precise timestamps to events, speakers, or musical elements.

2026-08-30 Method / Reward Modeling / Rubric-based reinforcement learning judging

Small Language Models as Judges for Rubric-Based Reinforcement Learning

This paper investigates whether small language models can serve as efficient and reliable judges for rubric-based reinforcement learning, where responses are evaluated against instance-specific criteria rather than exact-match verifiers.

2026-08-30 Evaluation / Bias Evaluation / Bias persistence assessment in text-to-image models

ContextBias: Controlled Evaluation of Bias Persistence Under Context Shift in Text-to-Image Models

This paper introduces ContextBias, a controlled evaluation framework designed to assess whether role-linked visual associations in text-to-image models persist when prompted context shifts.

2026-08-30 Application / Human-AI Interaction / Collaborative web development platform

VibeJam: A User Study Platform for Web Development with Agents

VibeJam is a browser-based platform designed for online user studies of agentic web development, where people collaborate with AI systems to build websites.

2026-08-30 Evaluation / Model Evaluation / Multiple-choice scoring impact

Check The Scoreboard: An Analysis of Scoring Schemes on Multiple-Choice Evaluation

This paper examines how multiple-choice question answering benchmarks change when evaluated with alternative scoring schemes rather than standard number-right accuracy.

2026-08-29 Method / Physics-Informed Learning / Preconditioning for PDE solvers

SS-ESOAP: Self-Scaled Adaptive Preconditioning for Physics-Informed Learning

This paper introduces SS-eSOAP, a preconditioner for physics-informed neural network training that extends SOAP-style Kronecker-factored optimization with two additions: a scalar secant-energy correction and adaptive eigenbasis updates with variance-state downscaling.

2026-08-29 Method / 3D Reconstruction / Interactive 3D and 4D scene editing

Chat-Edit-3D++: Interactive 3D and 4D Scene Editing via Large Language Models

This paper presents CE3D++, an interactive framework for editing 3D and monocular 4D scenes through multi-round dialogue with a large language model.

2026-08-29 Evaluation / Speaker Attribution / Benchmarking speaker attribution abilities

HEAR Who Said What: Unlocking Speaker-Attributed Reasoning via Counterfactual Voice Grounding

This paper studies whether speech language models can correctly determine who said what in multi-speaker audio and reason over speaker identities.

2026-08-29 Method / Reinforcement Learning / Reinforcement fine-tuning techniques

Dynamic Important Example Mining for Reinforcement Finetuning

This paper studies data utilization in reinforcement fine-tuning (RFT) for large language and vision-language models, arguing that most prior data-centric methods treat sample importance as fixed even though policy learning is non-stationary.

2026-08-29 Method / World Model / Navigation action modeling

Hydra: A Navigation World Action Model with Discrete Latent Planning and Continuous Flow-Matching Execution

Hydra is a navigation world action model designed to address the mismatch between generative world models and external planners that makes real-time control difficult.

2026-08-28 Method / Generative Modeling / Video-based geometric learning framework

Video Generative Models as Geometry Learner

This paper proposes GeoNeXt, a unified framework for monocular geometry estimation that repurposes a pretrained video generative model for depth and surface normal prediction.

2026-08-28 Method / Motion Control / Framework separating motion and spatial mapping

RASA: Disentangled Spatial-Motional Priors for Cross-Identity Character Animation

This paper studies cross-identity character animation, where a reference image provides the target identity and a driving video provides the motion, under challenging mismatches in position, scale, and skeletal proportions.

2026-08-28 Method / Video Grounding / Spatio-temporal grounding with decoding

Locate Anything in Videos: Rethinking Efficient Generative Spatio-Temporal Video Grounding

This paper studies spatio-temporal video grounding (STVG) in multimodal large language models, targeting the inefficiency of autoregressively generating dense bounding-box trajectories across time.

2026-08-28 Method / Dialogue Systems / Multilingual culturally grounded assistant simulation

CultureConverse: A Multilingual Multi-turn Simulation Harness for Culturally Grounded Assistance in East and Southeast Asia

CultureConverse is a multilingual simulation and evaluation harness for culturally grounded assistant dialogue across East and Southeast Asia.

2026-08-28 Method / 3D Model Refinement / Selective repair via region selection and target operations

InstructMesh: Selective Refinement of Generative 3D Models for Fabrication

InstructMesh is an interactive post-generation tool designed to repair fabrication-relevant flaws in AI-generated 3D models through user region selection and targeted latent-space editing.

2026-08-27 Task / Surface Detection / Glass surface detection in scenes

Glass Surface Detection Grounded in 3D Visual Geometry

This paper reframes glass surface detection as a 3D visual geometry-grounded task rather than relying predominantly on 2D appearance cues.

2026-08-27 Evaluation / Benchmarking / Probabilistic alignment evaluation

PAWBench: How Far Are We from Probabilistically Aligned World Modeling?

This paper argues that evaluating video generators as visual world models requires assessing the distribution of possible futures under a fixed observation and action, rather than judging only single-rollout plausibility.

2026-08-27 Method / Reinforcement Learning / Policy optimization via rollout agreement

TTPO: Test-Time Policy Optimization

This paper introduces Test-Time Policy Optimization (TTPO), a label-free test-time training method for mathematical reasoning models.

2026-08-27 Task / Backdoor Injection / Trapping failure modes in VLA models

TrapVLA: Trapping Vision-Language-Action Models in Configured Failure Modes

This paper studies a new backdoor attack setting for Vision-Language-Action models called Configured Failure Trapping, where a stealthy textual trigger causes a robot to fail in a specific, preconfigured way rather than simply fail arbitrarily.

2026-08-27 Method / Self-Distillation / On-policy distillation framework

Self-OPD: On-Policy Distillation for Flow Matching Models without Teacher

This paper presents Self-OPD, a teacher-free on-policy distillation framework for flow matching models.

2026-08-26 Method / Memory Networks / Dual-brain memory architecture

VoiceMem: Streaming Dual-Brain Memory for Real-Time Interaction

VoiceMem is a memory architecture for real-time spoken interaction that separates memory into a factual "left brain" and an affective/persona-oriented "right brain," connected through streaming memory I/O.

2026-08-26 Method / Agent / Task-adaptive agent harness synthesis

JIT-Agent: Scaling Harness Intelligence via Just-in-Time Harness Evolution

This paper presents JIT-Agent, a model trained to generate task-adaptive agent harnesses on demand rather than relying on a fixed, manually engineered scaffold.

2026-08-26 Evaluation / Instruction Following Evaluation / Multimodal LLMs on video tasks

Video-IFBench: Evaluating Instruction Following of Multimodal LLMs in Video Understanding Scenarios

Video-IFBench is a benchmark designed to evaluate how well multimodal large language models follow instructions in video understanding settings, rather than only measuring answer correctness.

2026-08-26 Method / Reinforcement Learning / Visuals-based reinforcement learning approach

V-Rubrics: Visual Faithfulness via Rubric-Based Reinforcement Learning

This paper studies visual faithfulness in vision-language model post-training and frames the problem as a credit-assignment issue rather than only an evaluation issue.

2026-08-26 Method / Video Action Modeling / In-context causal modeling from videos

Zero-WAM: In-Context World-Action Modeling from Human Videos for Open-Ended Task Generalization

This paper frames zero-shot cross-task robotic manipulation as an in-context learning problem, where an unseen task is specified at deployment time by a human demonstration video rather than only by language.

2026-08-25 Method / Model Safety / Step-level tool action checking

StepGuard: Learning Step-Level Guardrails with Scalable Supervision and Safety-Utility Balancing

StepGuard is a 4B step-level guard model for LLM agents that supports both pre-execution checking of candidate tool actions and post-hoc auditing of completed trajectories.