LoRA, PEFT, instruction tuning and domain adaptation — adapting a pretrained model without paying to train one.
arXiv:2608.23830v1 Announce Type: cross Abstract: RL has emerged as a powerful paradigm for enhancing the instruction following capabilities of LLMs. While existing training recipes achieve substanti...
arXiv:2608.24727v1 Announce Type: cross Abstract: EEG foundation models pretrained via self-supervised learning promise transferable representations, but their generalization remains limited, especia...
SketchGuard is a Byzantine‑robust decentralized federated learning method that separates neighbor screening from model aggregation by using a Count Sketch representation. The approach mitigates a vulnerability where an adaptive adversary can hide large perturbations in the sketch’s null space, by adopting a commit‑then‑sketch protocol that ensures the sketch seed is chosen only after model commitment. The authors prove convergence in both convex and non‑convex settings, demonstrate that SketchGuard achieves state‑of‑the‑art robustness against six attacks—including the adaptive null‑space attack—across various network topologies and data heterogeneity, while reducing per‑neighbor communication to a model‑dimension‑independent size.
The paper investigates how knowledge distillation (KD) applied at intermediate layers of a neural network can affect overfitting and model performance. While traditional KD focuses on the final output, this study explores block‑wise KD across eleven datasets, finding that on standard datasets the last block suffices, but on fine‑grained, data‑scarce settings intermediate supervision significantly improves accuracy. The authors also analyze optimal supervision granularity using attention maps, Centered Kernel Alignment, and Grad‑CAM, and examine teacher‑student fine‑tuning strategies.
AgentWorld is a simulation framework that evaluates agentic information retrieval by incorporating diverse user personalities based on the Big Five (OCEAN) traits, stateful tool-use environments, and a pass$^k$ consistency metric with structured fault classification and partial-credit scoring. It includes a risk analyzer that uses Monte‑Carlo rollouts and advanced scoring methods to quantify trajectory brittleness and attack attribution. Experiments with conversational analytics, customer‑support agents, and adversarial stress‑testing demonstrate that personality variation reveals failure modes hidden by uniform testing, such as cross‑domain leakage, contextual drift, and significant quality gaps across personas.
RePolicy is a reinforcement learning approach designed to invoke safety policies for language model agents by evaluating entire execution trajectories within context-dependent policy libraries. It generates policy-grounded rationales and safety judgments, and is initialized with the PolicyTraj-20K dataset before fine-tuning via GRPO with verifiable rewards and policy-context perturbation. Experiments on six safety benchmarks demonstrate strong safety-detection performance and robust policy invocation across varying contexts.
The paper introduces Hierarchical Skill Retrieval (HSR), a framework that decomposes a target manipulation task into candidate skill sequences and evaluates each plan for semantic plausibility and skill reliability. HSR combines subtask-level language retrieval with behavior-feature reranking to select demonstrations that are both relevant and compatible with the target task, followed by a two-stage pretraining and finetuning pipeline for policy adaptation. Experiments on the LIBERO benchmark and real-world robot tasks show that HSR improves average success rates by 10.3% and 21.3% over the strongest baseline, demonstrating the effectiveness of structured skill-level retrieval for data-efficient Vision‑Language‑Action adaptation.
arXiv:2506.12006v4 Announce Type: replace-cross Abstract: The cross-Modality Domain Adaptation (crossMoDA) challenge series, initiated in 2021 in conjunction with the International Conference on Medi...
The paper introduces ICI-Time, a framework that casts time series forecasting as a visual inpainting problem. By converting series into area‑chart images, it enables pre‑trained vision transformers to perform forecasting through in‑context learning without fine‑tuning or new temporal architectures. Experiments on epidemiology, meteorology, and power systems show competitive performance and strong adaptability in low‑data scenarios.
The paper presents a method for generating cardiac magnetic resonance (CMR) images conditioned on patient metadata using a pretrained latent diffusion model. By encoding structured clinical data and slice position as textual prompts and applying Metadata‑Free Classifier‑Free Guidance, Contrastive Batching, and Inverse‑Frequency Sampling, the authors improve the fidelity of synthetic images, achieving a 57% reduction in Fréchet Inception Distance compared to a baseline without these strategies. Evaluation on 59,058 UK Biobank CMR scans shows better distributional realism and subgroup alignment, though disease‑specific conditioning remains challenging.
arXiv:2608.23794v1 Announce Type: new Abstract: Mixture-of-Experts (MoE) scales language models by routing each input through a small set of independently parameterized experts. We show that copying...
PROOF-Gen is a method that improves distillation of tool‑calling models by recovering successful trajectories from teacher failures. It uses per‑scenario prompt optimization to generate corrective guidance that steers the teacher to a passing trajectory, then removes this guidance before training so the student learns from clean demonstrations. On τ2‑bench, PROOF-Gen recovers 93% of failed scenarios, boosting Qwen3‑4B‑Instruct‑2507’s Pass^1 from 0.132 to 0.529 and improving Gemma 4 E4B‑it by 7.2pp on BFCL v4 multi‑turn, while also raising deployed on‑device model performance by up to 5.0pp across response‑quality metrics.
arXiv:2603.02041v2 Announce Type: replace-cross Abstract: Large language models (LLMs) are predominantly trained on English-centric data, resulting in uneven performance for smaller languages. We stu...
arXiv:2608.24354v1 Announce Type: cross Abstract: MLLMs are increasingly deployed in user-facing applications, yet they inherit backdoor risks from the pipelines used to construct them: triggers may...
arXiv:2608.24364v1 Announce Type: new Abstract: Self-supervised pretraining enables transferable representations for medical imaging, yet most CT encoders remain biased toward coarse semantic underst...
arXiv:2608.23903v1 Announce Type: new Abstract: Visual foundation models are commonly adapted under the assumption that the appearance of incoming data may change while the semantic meaning of the pr...
arXiv:2608.23869v1 Announce Type: new Abstract: While state-of-the-art generative models produce high-fidelity 3D meshes, these outputs lack the physical properties required for interactive simulatio...
CoMMa is a decentralized large language model (LLM) agent framework designed for oncology decision support. It allows specialists to work on partitioned clinical data streams, enforcing data decentralization and enhancing role specialization through agent-specific fine‑tuning. The framework introduces a contribution‑aware aggregation mechanism that replaces stochastic reasoning with deterministic embedding projections, enabling explicit credit assignment and a stable, interpretable decision pathway.
QiMeng-ChipV-RTL is a multi‑agent framework that tackles the challenges of generating Register‑Transfer Level (RTL) code for industrial IP‑level hardware design. By partitioning long design documents into short, localized tasks and using hierarchical planning, localized code generation, interface‑consistent merging, and AST‑guided debugging, it scales to complex specifications. Experiments on the RealBench benchmark show ChipV-RTL achieves a 45.0% pass rate, outperforming state‑of‑the‑art LLMs and agents which reach only 21.6%.
The paper introduces Joint Distribution Alignment for Universal Domain Adaptation (JAUA), a new algorithm designed for scenarios where source and target domains have differing label spaces. It provides a theoretical upper bound on generalization error for Universal Domain Adaptation and proposes aligning joint distributions using Chi‑Square divergence, complemented by a progressive pseudo‑labeling strategy. Experiments on six public image datasets show JAUA outperforms existing methods in handling Universal Domain Adaptation challenges.