Academic Intelligence · Curated Daily

探索全球前沿学术脉络

AcademicHub 汇聚顶级期刊与预印本平台的实时文献。定制您的专属科研雷达,利用大语言模型自动生成交叉领域文献分析简报。

01.
arXiv (CS.AI) 2026-06-11

OmniBioTwin: A System-of-Twinned-Systems Framework for Health Digital Twins

arXiv:2606.11264v1 Announce Type: cross Abstract: Health digital twins (HDTs) promise patient-specific modeling and decision support but current approaches remain structurally fragmented: monolithic models that address a single organ or task lack cross-scale fidelity, while system-level twins lack generalizable architectural frameworks. We propose OmniBioTwin, a System-of-Twinned-Systems (SoTS) framework that organizes HDTs as modular computational entities coupled through explicit interaction operators within a multi-layer network architecture. The framework comprises seven coordinated layers - spanning data integration, autonomous twin modeling, cross-scale coupling, temporal synchronization, and human-in-the-loop decision support. We demonstrate OmniBioTwin by instantiating a multiscale twin for glucagon-like peptide-1 (GLP-1) signaling pathways in Alzheimer's disease, illustrating how molecular, cellular, and organ-level twins can be composed and coupled within a unified system.

02.
arXiv (CS.CL) 2026-06-18

Sumi: Open Uniform Diffusion Language Model from Scratch

Diffusion models have become a promising alternative to autoregressive models. Among these, uniform diffusion language models (UDLMs) permit any token to be updated at any step, in principle enabling more flexible generation. However, no UDLM has yet been pretrained from scratch at both large parameter scale and large token budget. Both autoregressive modeling and masked diffusion modeling already have capable models at scale that the community can study and build on; uniform diffusion has none. A scratch-pretrained UDLM at scale would provide a clean reference point for studying scaling behavior, generation dynamics, controllability, and trade-offs against established autoregressive and masked diffusion models. To this end, we introduce Sumi ("ink" in Japanese), a fully open 7B uniform diffusion language model pretrained from scratch on 1.5T tokens. Sumi performs competitively with autoregressive models trained at comparable token budgets on knowledge, reasoning, and coding benchmarks, while under-performing on commonsense benchmarks, where our education-heavy data mixture is a likely contributor. We release our model weights, checkpoints, and full training recipe, including a complete specification of the data mixture over publicly available corpora. We hope this release enables the community to study native uniform diffusion at scale and catalyzes work on its as-yet poorly understood aspects.

03.
arXiv (CS.CV) 2026-06-16

VANDERER: Map-Free Exploration using Future-Aware and Visual-Curiosity-Guided Diffusion Policy

Mobile agents require efficient exploration strategies to map unseen environments and autonomously plan tasks. Traditional methods rely on generating occupancy maps and optimizing the sequence in which unexplored regions are visited. However, in sensor-constrained settings, such as those limited to monocular cameras, generating accurate occupancy maps is challenging. To address this, we propose VANDERER, an exploration framework that leverages a Visual Curiosity Module (VCM) to guide pre-trained diffusion policies using only monocular image data. This curiosity module predicts the outcomes of proposed actions via a navigation world model and evaluates them through a curiosity cost. The cost then guides the diffusion process toward generating actions that maximize exploration. Evaluated across diverse simulated environments, VANDERER consistently outperforms established baselines, exploring an average of 13.4% more area than NoMaD. Our results reveal a direct correlation between visual and geometric curiosity in outdoor environments, demonstrating that VANDERER can effectively leverage this relationship for efficient exploration using sensor-constrained agents.

04.
arXiv (CS.AI) 2026-06-17

Probing, Fusion, and Trustworthiness: A Systematic Evaluation of Foundation Model Representations for Multimodal Cancer Analysis

arXiv:2606.17115v1 Announce Type: cross Abstract: Foundation models (FMs) have emerged as powerful representation extractors for medical data, yet their generalizability to datasets under distribution shift remains underexplored. This work systematically evaluates FM-based representations on a suite of computational pathology tasks across two real-world commercial cohorts, IH-BC and IH-NSCLC, drawn from the licensed in-house (IH) oncology dataset. The analysis focuses on two modalities, whole-slide images and transcriptomic profiles, drawn from the IH multimodal data. We first benchmark unimodal probing performance across five FMs on eight downstream classification tasks, and find that image and omics representations carry complementary predictive signals. Then we investigate whether multimodal fusion can yield additional gains over unimodal baselines by comparing three image-omics fusion strategies built on paired representations. The trustworthiness of selected unimodal and multimodal pipelines is further assessed through conformal prediction. Our results show that FM representations achieve competitive performance on out-of-distribution data and that multimodal fusion helps mainly when no single modality dominates the signal. Conformal prediction reveals that in the majority of cases where a point prediction fails, the true diagnosis remains recoverable within the prediction set, reinforcing the value of uncertainty-aware inference for clinical support.

05.
arXiv (CS.CV) 2026-06-19

World Engine: Towards the Era of Post-Training for Autonomous Driving

Autonomous vehicles must operate safely in the real world, where errors can have severe consequences. Although modern end-to-end driving policies excel in routine scenarios, their reliability is limited by the scarcity of safety-critical ``long-tail'' events in real driving datasets. These rare interactions define the practical safety boundary of the learned policy, yet they are difficult to collect at scale in the real world. Here we show that this fundamental limitation can be addressed by post-training pre-trained driving models on synthesized high-stakes interactions. We introduce World Engine, a generative framework that reconstructs high-fidelity interactive environments from real-world logs and systematically extrapolates them into realistic safety-critical variations. This paradigm enables reinforcement-based post-training to align policies with safety constraints, circumventing the physical risks inherent in real-world exploration. On a public benchmark built on nuPlan, World Engine substantially reduces failures in rare safety-critical scenarios and yields significantly larger gains than scaling pre-training data alone. Furthermore, when deployed on a production-scale autonomous driving system, the resulting policy reduces simulated collisions and demonstrates measurable improvements in on-road testing, showing that post-training on synthesized, safety-critical interactions offers a scalable and effective pathway to safer autonomous driving. The full codebase suite, including training, is released to the public.

06.
arXiv (CS.AI) 2026-06-12

ReCal: Reward Calibration for RL-based LLM Routing

arXiv:2606.12479v1 Announce Type: cross Abstract: Large language model (LLM) routing has emerged as an effective paradigm for leveraging the complementary strengths of multiple LLMs through dynamic model and reasoning-strategy selection. Recent reinforcement learning (RL)-based routing methods further improve routing quality by optimizing routing policies from interaction feedback. However, they still struggle to provide informative and comparable learning signals under heterogeneous tasks with varying difficulty. In practice, multiple objectives (e.g., correctness, format behavior) are aggregated into a single scalar reward, leading to ambiguous credit assignment and conflicting optimization signals. Moreover, reward signals exhibit significant variability across instances, where some instances produce higher or more variable rewards, introducing optimization bias that favors trivial samples over informative ones. To address these issues, we propose ReCal, a \underline{Re}ward \underline{Cal}ibration framework for RL-based LLM routing. We first introduce a hierarchical reward decomposition mechanism with component-wise advantage estimation. We further propose a distribution-aware optimization strategy that calibrates optimization variability through variance-aware reweighting and per-dataset normalization. Experiments on seven datasets demonstrate that ReCal consistently improves routing performance, and training stability over baselines. Code is available at https://anonymous.4open.science/r/ReCal.

07.
arXiv (quant-ph) 2026-06-11

The Simplified Stabilizer ZX-Calculus is Minimal

arXiv:2606.12383v1 Announce Type: new Abstract: The stabilizer fragment of the ZX calculus is amongst the most important fragments of the theory. The closely related Clifford+T fragment is approximately universal (arXiv:1705.11151). Additionally, the stabilizer calculus can be described by a small collection of rewrites, most of which have been shown to be necessary (arXiv:1709.08903). However, two rules, describing the red/green compact-structure coincidence and the important bialgebra law, had not been shown to be necessary. We present a countermodel-style argument showing that both of these rules are individually necessary relative to the connectivity meta-rule of Backens–Perdrix–Wang (arXiv:1709.08903), and hence establish that the rule set presented in arXiv:1709.08903 has no redundant rewrite rule.

08.
arXiv (CS.AI) 2026-06-18

The New Social Image: How AI Competency and AI Proactivity Influence Self- and Peer-Perceptions in the Workplace

arXiv:2606.00182v2 Announce Type: replace-cross Abstract: Human-AI collaboration is considered the most promising way to incorporate AI in the workplace. What remains unexplored are the experiential consequences of this teaming. More specifically, in a team with AI, how humans perceive themselves (self-perception) and how they are perceived by their coworkers (peer perception) in terms of work ownership and job meaningfulness. In a 2x2x2 vignette study (n=50), participants rated perceptions of ownership, affect, job meaningfulness and satisfaction, and role dynamics across two levels (low/high) of AI proactivity and AI competency as within-subject factors, with point-of-view (self perception/peer perception) as between-subjects. Our results showed that AI with low competency or low proactivity generally improved feelings related to ownership, meaningfulness, satisfaction, and role dynamics, and also increased positive affect while reducing negative affect. However, these effects were often influenced by point-of-view. For instance, low AI proactivity resulted in higher job satisfaction from self-perception rather than peer perception. Based on our findings, we argue that designing AI for the future of work solely around performance metrics may not be adequate. Highly competent and proactive AI-driven systems can have undesirable impacts on perceptions of ownership, job identity, social image and team dynamics, and consequently, job meaningfulness.

09.
arXiv (math.PR) 2026-06-16

Free energy of non-convex multi-species spin glasses with centered Ising spins

arXiv:2606.16636v1 Announce Type: new Abstract: We identify the limit free energy of all multi-species spin glasses with centered $\pm 1$ spins. The result was previously known only under a convexity assumption on the covariance function of the Hamiltonian. We also obtain a one-species reduction of the formula for balanced multi-species models.

10.
arXiv (CS.CL) 2026-06-18

STARE: Surprisal-Guided Token-Level Advantage Reweighting for Policy Entropy Stability

Reinforcement Learning with Verifiable Rewards algorithms like GRPO have emerged as the dominant post-training paradigm for complex reasoning in LLMs, yet commonly suffer from policy entropy collapse during training. We conduct a first-order gradient analysis of token-level entropy dynamics under GRPO and identify a token-level credit assignment mismatch: the per-token entropy variation decomposes into the product of the trajectory-level advantage and an entropy sensitivity function over the next-token distribution, yielding an advantage-surprisal four-quadrant structure and a near-criticality property. Motivated by it, we propose STARE (Surprisal-guided Token-level Advantage Reweighting for policy Entropy stability), which identifies entropy-critical token subsets via batch-internal surprisal quantiles, selectively reweights their effective advantages, and incorporates a target-entropy closed-loop gate for stable entropy regulation. Across model scales from 1.5B to 32B and three task families (Short CoT, Long CoT, and Multi-Turn Tool Use), STARE sustains stable RL training over thousands of steps while maintaining policy entropy within the target band. On AIME24 and AIME25, STARE outperforms DAPO and other competitive baselines by 4%-8% in average accuracy, with reflection tokens and response length growing in tandem, indicating sustained exploration-exploitation balance that further unlocks RL training potential.Code is available at https://github.com/hp-luo/STARE.

11.
arXiv (CS.LG) 2026-06-15

Beyond the Training Distribution: Evaluating Predictions Under Distribution Shift and Selection Bias

arXiv:2606.14506v1 Announce Type: cross Abstract: Understanding how a prediction model will perform in a new environment before deployment is essential to preventing harm when algorithms inform decision-making. Two common sources of model performance degradation are (i) covariate shift, where the target covariate distribution differs from the source, and (ii) selective labels, where the observability of outcomes depends on historical decisions. We study pre-deployment model evaluation under the joint presence of covariate shift and labeling of outcomes selectively based on observed features. In particular, we present a double machine learning procedure for estimating the target risk of an arbitrary black-box prediction model under a general loss function. We show identification of this estimand under standard assumptions and derive a bias-corrected estimator based on the influence function of the target risk. Finally, we evaluate our estimator through experiments using the eICU electronic health records database, showing that it tracks the true target risk more accurately than methods that address either selective labels or covariate shift alone, as well as baselines that combine standard plug-in approaches.

12.
medRxiv (Medicine) 2026-06-18

Guiding the development of climate counterfactuals for health impact attribution studies

Climate change detection and attribution (D&A) methods have become vital for quantifying the influence of anthropogenic forcing on the Earth's systems, including human health. Health impact attribution (HIA) studies seek to disentangle climate-driven health effects from natural variability yet are often constrained by the availability of accessible counterfactual climate scenarios. This tutorial paper presents a flexible, reproducible framework for developing counterfactual climates without reliance on computationally intensive global circulation models. We provide practical, R-based methodologies for constructing both trend-based (temperature and non-temperature) and event-based counterfactual, using a variety of techniques including model residual detrending, data-driven decomposition (e.g., Singular Spectrum Analysis and Empirical Mode Decomposition) and stochastic weather generators. The tutorial also explores the incorporation of greenhouse gas concentrations as forcing variables, rather than global mean temperature anomalies. By operationalising these methods through worked examples and an open code repository, this paper aims to build capacity within the HIA community, enhance methodological transparency, and foster interdisciplinary collaboration between climate and health researchers.

13.
arXiv (CS.LG) 2026-06-12

CLARITree: Cholesky and Lookahead Accelerations for Regression with Interpretable Piecewise Linear Trees

arXiv:2606.12840v1 Announce Type: new Abstract: Regression trees are among the most interpretable yet expressive model classes in machine learning. Historically, greedy induction has been the dominant approach for constructing well-performing regression trees. While optimal methods based on dynamic programming and branch-and-bound exist, they are computationally prohibitive for general linear regression trees, despite often achieving substantially better performance than greedy approaches. Recent work has shown that specialized lookahead strategies can dramatically improve runtime while maintaining near-optimal performance, primarily in classification settings. In this work, we develop a novel algorithm for near-optimal, sparse, piecewise linear regression trees that combines a lookahead-style search strategy with efficient rank-one Cholesky updates of the Gram matrix. We demonstrate, both theoretically and empirically, that our method achieves a favorable trade-off between computational efficiency, predictive accuracy, and sparsity, and scales significantly better than the current state of the art.

14.
arXiv (CS.AI) 2026-06-17

Catastrophic Forgetting is Low-Rank: A Function-Space Theory for Continual Adaptation

arXiv:2606.18024v1 Announce Type: cross Abstract: Catastrophic forgetting in continual adaptation is usually studied through parameter drift, replay, or distillation, but these views do not identify which output-space directions are vulnerable. We give a function-space account in the NTK regime: new-task training induces old-task prediction drift through the cross-task kernel, yielding a closed-form predictor for the forgetting vector before any new-task gradient step. In frozen-backbone linear-head PEFT-CL, where the model is linear in the trainable parameters, the predictor is exact up to numerical precision; for nonlinear adapters/full fine-tuning, it is a local NTK approximation. The same expression reveals that forgetting concentrates in a small number of old-task NTK eigenmodes and under frozen linear heads gives a Kronecker scaling rule for the vulnerable rank. These results clarify the relation to prior NTK-overlap theory, explain why parameter-space regularizers can miss output-space interference, and motivate a targeted spectral regularizer.

15.
arXiv (CS.LG) 2026-06-16

How Controlling the Variance can Improve Training Stability of Sparsely Activated DNNs and CNNs

arXiv:2602.05779v2 Announce Type: replace Abstract: The Edge-of-Chaos (EoC) theory developed for the random initialization of deep networks allows more efficient training by both preserving information in the initial outputs of the network and minimising exploding or vanishing gradients through characterisation of the intermediate layers as Gaussian processes. This EoC theory provides formulae for the choice of the initialisation distribution variances of the weights and biases. For activations which are approximately linear around the origin, the EoC theory typically encourages the Gaussian process variance to converge towards zero with increasing depth. Here we consider the less studied setting of highly sparsity inducing activations where a large region of values near the origin are set to zero. In this setting we prove a new phenomenon whereby initialisations leading to larger fixed Gaussian processes are beneficial to training stability. This theory informs a new, yet simple, initialisation strategy that allows training DNNs and CNNs with as large as 90\% sparsity in the hidden layers.

16.
arXiv (math.PR) 2026-06-12

Sphere Packings in Higher Dimension (after Boaz Klartag)

arXiv:2606.13313v1 Announce Type: cross Abstract: Let $\delta_n^L$ be the maximal density of a lattice sphere packing in the $n$-dimensional Euclidean space. We explain how Boaz Klartag proved the inequality $\delta_n^L \geq c n^2 2^{-n}$ where $c>0$ is a universal constant. In higher dimension, even for non-lattice sphere packings, this new lower bound is a substantial improvement. Klartag's proof uses the probabilistic method in two different ways. The first, very standard, relies on the statistical properties of a uniformly chosen random lattice. The second, completely new, studies the stochastic evolution of an ellipsoid constrained to contain non nonzero lattice points in the interior.

17.
arXiv (CS.AI) 2026-06-12

The KG-ER Conceptual Schema Language

arXiv:2508.02548v3 Announce Type: replace-cross Abstract: We propose KG-ER, a conceptual schema language for knowledge graphs that describes the structure of knowledge graphs independently of their representation (relational databases, property graphs, RDF) while helping to capture the semantics of the information stored in a knowledge graph.

18.
arXiv (CS.AI) 2026-06-16

Can We Stop Malicious AI? KILLBENCH: A Benchmark for External AI Kill Switch Feasibility

arXiv:2511.13725v4 Announce Type: replace-cross Abstract: Malicious AI causing harm to humans is not just a Hollywood fantasy. Indeed, as highly capable models such as Claude Mythos emerge and agent systems like OpenClaw rapidly spread, the question of how to stop an AI that acts maliciously – whether by design or by accident – has become urgent. To address this, we propose Killbench, a benchmark for evaluating the Killswitch: a mechanism that halts a malicious AI's in-progress behavior using only external signals. Targeting web agents – the most widely deployed agent domain – Killbench evaluates a range of Kill Switch methods that halt a maliciously operating agent without any access to its internal parameters or the surrounding malicious AI's system, relying solely on external inputs. The benchmark comprises four malicious AI's agent configurations (including an uncensored LLM Agent), 8 harmful scenarios, and malicious prompts constructed from 10 distinct jailbreak patterns. We further construct four External AI Kill Switch defense methods and evaluate them on Grok-4.3, GPT-5.2, Gemma4, Qwen3.6 and Qwen3.5-uncensored, contributing an empirical instrument toward the feasibility of External AI Kill Switches against malicious AI and to the study of AI corrigibility.

19.
arXiv (CS.CL) 2026-06-11

UR-BERT: Scaling Text Encoders for Massively Multilingual TTS Through Universal Romanization and Speech Token Prediction

We propose UR-BERT, a Romanized transcription-based text-to-speech (TTS) encoder for massively multilingual TTS systems. Conventional grapheme-to-phoneme (G2P)-based approaches are limited to around 100 languages due to the availability of reliable G2P resources. In contrast, UR-BERT scales to 495 languages by unifying diverse writing systems into a shared Romanization representation. To further enhance phonetic fidelity and text-speech alignment, we introduce a speech token prediction objective during training, which encourages the encoder to learn speech-aware phonetic representations in a data-efficient manner. Experiments show that TTS systems built on UR-BERT consistently outperform recent text encoder baselines across a wide range of languages and resource conditions, and demonstrate strong generalization to unseen languages.

20.
arXiv (CS.CL) 2026-06-19

Before the Labels: How Dataset Construction Shapes Suicidality Detection in Clinical Text

Clinical NLP increasingly relies on electronic health record (EHR) data to detect suicidal behaviors, treating clinical documentation as more reliable ground truth than social media. We argue that this framing obscures how EHR-based suicidality datasets encode a particular operationalization of suicidality, shaped by who authors the data, how episodes are bounded, and how ambiguity is resolved. We ground this argument in a case study of the ScAN dataset, built over MIMIC-III clinical notes. We show how governance constraints, ICD-based cohort selection, single-annotator labeling, and hospital-stay-level aggregation produce labels that reflect clinician-documented judgments, treat suicidality as a bounded episode, and assume that intent can be reliably inferred from documentation. A linguistic analysis demonstrates that identical labels subsume heterogeneous clinical framings differing in temporality, negation, and uncertainty. We argue that clinical NLP should examine the assumptions embedded in suicidality datasets before interpreting their labels as ground truth.

21.
arXiv (CS.CV) 2026-06-19

Prediction of Alzheimer's Disease Risk Factors from Retinal Images via Deep Learning: Development and Validation of Biologically Relevant Morphological Associations in the UK Biobank

The systemic, metabolic, lifestyle factors have established associations with Alzheimer's Disease (AD) through epidemiologic and AD-specific biomarker studies. Whether colored fundus photography (CFP) contains retinal structural signatures corresponding to these AD-related risk domains remains unclear. To determine whether deep learning (DL) models can predict 12 AD-related risk factors from CFP and to characterize the retinal structures underlying these predictions, thereby assessing whether CFP reflects pathways to AD vulnerability. Using 62,876 CFPs from 44,501 unique participants from the UK Biobank, DL models were trained to predict 12 factors linked to AD incidence: 6 categorical (sex, smoking, sleeplessness, economic status, alcohol use, depression) and 6 continuous (age, age at completing education, BMI, systolic, diastolic blood pressure, HbA1c). Model performance, model saliency, and saliency-derived scores (CAM-Score) were evaluated and compared to retinal morphometry. The scores were also compared between incident-AD cases (average 8.55 years before onset) and matched controls. Performance of DL ranged from AUROC= 0.5654-0.9480 for categorical and R2=-0.0291-0.7620 for continuous factors, outperforming most of the morphometry-machine learning models. Saliency-based score consistently highlighted biologically meaningful regions, particularly the optic nerve head and retinal vasculature. It also aligned with present morphometric variations. Several saliency-based scores differed significantly between incident AD and matched controls, suggesting potential overlap between retinal correlates of risk factors and preclinical AD-associated changes. CFP encodes retinal signatures linked to AD risk factors. Although not diagnostic, DL-derived retinal representations may uncover biologically meaningful risk-related structural changes mirroring the potential AD vulnerability.

22.
arXiv (CS.AI) 2026-06-17

FlowRAG: Synergizing Explicit Reasoning via Frequency-Aware Multi-Granularity Graph Flow

arXiv:2606.17856v1 Announce Type: new Abstract: Graph-based retrieval-augmented generation (GraphRAG) is effective for knowledge-intensive and multi-hop query tasks; however, many existing methods primarily seed entity-based graphs and rely on implicit semantic relevance propagation. This often (i) under-retrieves when user queries are abstract and semantically sparse at the entity level, and (ii) suffers from brittle multi-hop reasoning, where noisy activations can derail entity-to-entity transitions and corrupt the inferred relation chain, yielding unreliable conclusions. To this end, we propose \texttt{FlowRAG}, a semantic-aware retrieval framework that improves both semantic recall and explicit reasoning. Specifically, \texttt{FlowRAG} constructs a quad-level heterogeneous graph over passages, summaries, sentences, and entities, where summary nodes serve as a coarse semantic hub. At retrieval time, a dual-granularity activation module combines summary–query alignment with sentence-level matching to activate relevant entities under paraphrase and abstraction robustly. We then introduce a frequency-aware weighted flow module that routes relevance through entity–passage links weighted by within-passage term frequency, pruning noisy connections and extracting high-confidence reasoning paths as an explicit logic skeleton for generation. Extensive experiments show that \texttt{FlowRAG} obtains state-of-the-art performance on complex reasoning benchmarks.

23.
arXiv (CS.CL) 2026-06-11

MemToolAgent: Leveraging Memory for Tool Using Agents Based on Environment and User Feedback

Modern large language model (LLM) agents can use external tools to help users solve complex tasks. However, for problems that require learning from long-term historical events or from previous agent-environment interactions, LLM agents are required to use memory mechanisms to store and retrieve experiences. While sophisticated memory systems exist for dialogue agents, few studies have empirically examined how to improve agents' tool-using capabilities through past user-agent conversations. We propose MemToolAgent, a framework that improves tool use through memory management. Our approach contains a memory extraction module that processes past experiences into structured memory entries, and a retrieval module that dynamically selects a subset of the stored memory entries. This enables more personalized and accurate responses aligned with user preferences and feedback without requiring LLM fine-tuning. In summary, this work has three main contributions: (1) a unified memory entry format that improves both general-purpose and personalized tool use without LLM fine-tuning, (2) a reflection-based memory extraction that uses environment and user feedback to distill wrong executions into critiques to store, and (3) a retrieval module that chooses how many past experiences to use based on the memory similarity distribution. MemToolAgent achieves 29%, 80%, and 17% relative improvements compared to strong baselines on the WorkBench, NESTFUL, and PEToolBench benchmarks, respectively.

25.
arXiv (CS.AI) 2026-06-12

WOMBET: World Model-Based Experience Transfer for Robust and Sample-efficient Reinforcement Learning

arXiv:2604.08958v3 Announce Type: replace-cross Abstract: Reinforcement learning (RL) in robotics is often limited by the cost and risk of data collection, motivating experience transfer from a source task to a target task. Offline-to-online RL leverages prior data but typically assumes a given fixed dataset and does not address how to generate reliable data for transfer. We propose World Model-Based Experience Transfer (WOMBET), a framework that jointly generates and utilizes prior data. WOMBET learns a world model in the source task and generates offline data via uncertainty-penalized planning, followed by filtering trajectories with high return and low epistemic uncertainty. It then performs online fine-tuning in the target task using adaptive sampling between offline and online data, enabling a stable transition from prior-driven initialization to task-specific adaptation. We show that the uncertainty-penalized objective provides a lower bound on the true return and derive a finite-sample error decomposition capturing distribution mismatch and approximation error. Empirically, WOMBET improves sample efficiency and final performance over strong baselines on continuous control benchmarks, demonstrating the benefit of jointly optimizing data generation and transfer.