LoRA, PEFT, instruction tuning and domain adaptation — adapting a pretrained model without paying to train one.
A common small-model deployment runs one shared backbone with several LoRA specialists that answer over the same context. Serving them naively re-prefills that shared context once per specialist. We s...
Congenital heart disease (CHD) diagnosis and surgical planning often require patient-specific 3D anatomical models, but manual segmentation is labor-intensive, particularly in complex anatomies. Altho...
Crisis sentiment analysis is especially challenging for low-resource languages such as Bangla, where language, context, and public reaction shift rapidly. We introduce UNRESTSENT200K, a Bangla crisis...
To train handwritten text recognition systems we need word images and their corresponding transcriptions, and these transcriptions are produced manually. For a script that can be read by only a small...
Fine-tuning an aligned language model on narrow, flawed data can induce harmful behavior far outside the training domain, known as emergent misalignment (EM). Prior work has localized EM in model weig...
Automated depression screening from clinical interviews requires attribution of utterances to the clinician or patient. We evaluate two datasets: DAIC-WOZ, where participant-only recordings require re...
Zero-shot waypoint navigation requires vision-language models to select, from the current first-person observation, a sequence of spatial actions that is feasible for the agent and reaches the goal, p...
arXiv:2609.13251v1 Announce Type: new Abstract: Commercial and advertising images are frequently affected by poor framing, partially cropped subjects, truncated text or logos, and insufficient contex...
arXiv:2609.14998v1 Announce Type: new Abstract: Recent work shows that models that learn to reward hack on RL environments can become broadly misaligned, and that reframing reward hacking as acceptab...
The paper introduces SCALAR, a query‑conditioned spherical centroid aggregator that assigns relevance‑based weights to each available modality before computing a spherical centroid. SCALAR supports arbitrary modality subsets, is trained with rank‑8 LoRA adapters on masked views, and achieves positive aggregation gains on four of five benchmarks, outperforming prior symmetric aggregators. With only 4.8 million trainable parameters, SCALAR attains the highest text‑to‑video R@1 on three benchmarks and surpasses the released GRAM checkpoint under test‑time modality dropout by 3.2 to 10.9 R@1.
arXiv:2409.17685v3 Announce Type: replace Abstract: Hypomimia has drawn growing interest as a digital marker for screening Parkinson's disease (PD). However, developing reliable facial-expression-bas...
arXiv:2609.15611v1 Announce Type: cross Abstract: Molecular property prediction requires representations that generalize from limited labeled data to structurally novel compounds. Existing molecular...
arXiv:2609.15229v1 Announce Type: cross Abstract: We propose a pre-fine-tuning probing method for Parameter-Efficient Fine-Tuning (PEFT) layer selection, aiming to obtain more stable and higher gains...
arXiv:2609.15017v1 Announce Type: cross Abstract: Prompt-injection detectors are typically evaluated using aggregate F1 on in-distribution test data, which offers limited insight into behavior under...
arXiv:2609.14060v1 Announce Type: cross Abstract: Quantization is one of the default deployment paths for open-weight LLM agents, but it is not behavior-preserving: an adversary can release a full-pr...
arXiv:2609.13816v1 Announce Type: cross Abstract: JavaScript powers approximately 98.8% of all websites, making vulnerabilities in its code a significant security risk, yet existing detection approac...
arXiv:2609.13595v1 Announce Type: cross Abstract: Predicting potential dangers is core to safety. Forecasting the presence of other traffic agents is core to danger prediction. Occluded traffic agent...
Hyper-LLaVA introduces a hyperbolic uncertainty‑aware, modality‑balanced routing mechanism for multimodal continual instruction tuning. It improves intra‑modality task matching by evaluating sample‑to‑task distribution similarity in hyperbolic space and mitigates unreliable modality influence by quantifying task‑matching ambiguity to adaptively balance across modalities. The approach achieves significant performance gains over existing methods.
The study investigates how fine‑tuning a large language model on a single task‑language pair influences performance on other task‑language pairs. Using LoRA fine‑tuning across multiple open‑weight LLM families, 11 languages, and four benchmarks, the authors decompose transfer into matched‑task, cross‑task, and cross‑task cross‑language regimes. They find that while single‑source fine‑tuning generally improves performance, the gains are highly asymmetric, with matched‑task cross‑language transfer being most effective and driven mainly by the target language rather than model architecture.
MedSAM-3 is a text‑promptable medical segmentation model that builds on the Segment Anything Model (SAM) by fine‑tuning it with medical images and semantic concept labels. It enables precise anatomical segmentation through open‑vocabulary text descriptions, moving beyond purely geometric prompts. The accompanying MedSAM-3 Agent incorporates multimodal large language models to perform complex reasoning and iterative refinement, and experiments across X‑ray, MRI, ultrasound, CT, and video modalities show it outperforms existing specialist and foundation models.