PREEN (Preset Regression via an Exemplar-based Editor Network) performs color style transfer by predicting a 29‑parameter preset for a deterministic image editor, rather than generating a raster image. It uses a style‑transfer teacher to create pseudo‑presets, then trains a CLIP‑conditioned regressor refined with LoRA and a CIELAB loss. On the PST50 benchmark, PREEN outperforms baselines in pixel‑ and perceptual‑fidelity, runs in sub‑second on‑device latency, and offers a single inference‑time strength knob that smoothly transitions from default to vivid presets.
By Beomjin Ahn
The paper introduces SimpleLoRIF, a scalable gradient‑embedding technique that pinpoints fine‑tuning examples whose gradients align with unsafe behavior in large language models. By applying this method, the authors create SafeTulu, a version of Tulu 3 with the 20 % of most unsafe‑aligned examples removed, which reduces GCG attack success by up to 22 % across several LLMs while largely preserving or improving capability. The study also highlights previously under‑studied safety‑degrading data sources such as tool‑execution failures and narrative arithmetic.
By Saket Reddy, Amro Abbas, Anshuman Suri, Ari Morcos, David Schwab, Pratyush Maini
arXiv:2509. 00924v2 Announce Type: replace Abstract: Transformers often transfer across related tasks with little fine-tuning; we study this question in noisy nonparametric regression for continuous maps $f,g:[0,1]^d\to\mathbb{R}^D$, representing the source and shifted tasks, using a single-query, single-head transformer $\mathcal{T}(\cdot)=V\operatorname{Softmax}(-\alpha(\cdot,K))$ with $\ell^{\infty}$-based alignment scores $\alpha$.
By Anastasis Kratsios, Tin Sum Cheng, Bum Jun Kim, Daniel Roy
DuplexTune is a framework that enables memory‑efficient adaptation of mixture‑of‑experts language models by replaying coupled hidden states through pretrained attention and expert modules. It introduces trainable width bridges that connect two residual streams to the backbone, with a bridge warm‑up phase before backbone adaptation while keeping router weights frozen. Experiments on Qwen1.5‑MoE‑A2.7B show that DuplexTune reduces peak memory from 64.8 GB to 38.8 GB and increases throughput from 19.9 to 25.1 samples/s, achieving competitive performance on MMLU, GSM8K, and MT‑Bench.
By Paul Raj, Simon Caldwell, Daniel Whitmore, Mutlu Mete
The paper introduces BUSTA, a method for training small language models to reason about binary execution. By fine‑tuning Qwen3‑1.7B and Nemotron‑3‑Nano with Group Relative Policy Optimization, the authors improve the models’ ability to predict input register values for a given x86‑64 basic block, raising accuracy from 73.2% to 85.3% and from 63.3% to 95.2% respectively. The fine‑tuned Nemotron‑3‑Nano achieves performance within four points of frontier models while using far fewer parameters, and the authors release both the fine‑tuned models and the training dataset.
By Ritika Poddar, Aidan Fennelly, Deirdre Murphy, Eric Wolos, Corre L. Steele, Michael Doyle
The paper introduces the first open Speech Language Model for Basque, built on the LLaMA‑Omni architecture and a Latxa backbone, aimed at low‑latency, end‑to‑end spoken interaction. To overcome data scarcity, the authors translated the English VoiceAssistant‑400K corpus into Basque and synthesized audio with a Basque TTS system. Evaluation shows the model achieves a content quality score of 3.86/5, a 9.94% word error rate for intelligibility, and a 3.07 UTMOS for audio quality, demonstrating the feasibility of advanced spoken interaction in a low‑resource language.
By Andoni Sudupe, Inma Hernaez, Ander Arriandiaga, Christoforos Souganidis, Ibon Saratxaga, Eva Navas
The study evaluates ten ImageNet‑pretrained deep learning models for classifying benign versus malignant breast cancer in the BreakHis histopathology dataset, using a patient‑level split. DenseNet121 achieved the highest performance with 99.4% accuracy and an AUC‑ROC of 0.999, outperforming all other models (p < 0.001). Explainability methods (Grad‑CAM, LIME, Integrated Gradients, LRP) consistently identified nuclei‑rich and glandular regions as key diagnostic features, demonstrating that transfer learning combined with explainable AI can produce accurate and interpretable breast cancer diagnostic tools.
By Sami Alnawali, Muhammad Atif Saeed, Abdul Hadi, Azka Atiq, Akhtar Jamil, Ahmad Din
SkipCat introduces a rank‑maximized low‑rank compression framework for large language models, combining intra‑layer shared low‑rank projections and block‑skipping techniques to reduce redundancy and computational overhead. By allowing higher retained ranks while maintaining the same compression rate, SkipCat achieves a 7% accuracy improvement on zero‑shot tasks compared to prior low‑rank methods, all without additional fine‑tuning.
By Yu-Chen Lu, Sheng-Feng Yu, Hui-Hsien Weng, Pei-Shuo Wang, Yu-Fang Hu, Liang Hung-Chun, Hung-Yueh Chiang, Kai-Chiang Wu
The paper introduces Dr. Post-Training, a framework that treats general training data as a data‑induced regularizer to prevent overfitting to scarce target objectives, rather than merely a pool for selection. It constructs a feasible set of model update directions from general data and projects target‑driven updates onto this set, unifying standard training and existing data‑selection methods as special cases along a bias‑variance spectrum. Experiments on SFT, RLHF, and RLVR demonstrate that these methods outperform both standard training and state‑of‑the‑art data‑selection baselines, especially when general and target data are highly misaligned, while system optimizations keep overhead minimal.
By Pingbang Hu, Xueshen Liu, Z. Morley Mao, Jiaqi W. Ma
The paper introduces StressSim, an intent‑preserving adversarial user simulator designed to stress‑test multi‑turn agents by generating challenging yet valid user interactions. StressSim optimises high‑level adversarial plans for each task instance using a genetic prompt‑optimisation loop that leverages rollout traces, task‑completion signals, and validity feedback to reduce task success while keeping the user’s goal unchanged. Experiments on MultiWOZ, retail, and airline domains show that StressSim reveals significant robustness gaps in leading LLM agents—reducing tau‑bench pass^4 scores by up to 69%—while maintaining realistic conversation quality and goal preservation with a modest optimisation budget.
By Atri Vivek Sharma, Brian Formento, Alessio Lomuscio
The paper introduces Dual-Path Spatiotemporal Contrastive Learning (DP-STCL) for cross-subject EEG emotion recognition, using two complementary encoders—Temporal-to-Spatial and Spatial-to-Temporal—to model EEG data under different integration orders. It enforces consistency of high-level emotion representations across these paths and across subjects through three complementary relations, enhancing stability and discriminability. A LoRA-based few-shot adaptation strategy is added to handle limited labeled data from target subjects, and experiments on SEED, SEED-IV, SEED-V, and FACED datasets show DP-STCL outperforming existing methods.
By Zehui Xiao, Yi Zheng, Guixing Fu, Wenkai Lu, Lingfei He, Tao Tan, Yue Liu, Silun Wang, Junhai Xu, Mengting Liu
RoboJev-27B is an open‑weight vision‑language model that predicts whether a robot has successfully completed a manipulation task by analyzing a task instruction, ten frames of the attempt, and optional auxiliary camera frames. Trained with a rank‑32 LoRA adapter on 52,598 episodes, it achieves 95.2% balanced accuracy and 0.993 AUROC on held‑out episodes of the same robot model, and maintains strong performance on unseen labs, robots, and community‑collected data. The model runs in 0.26 seconds per episode on an H100 GPU and is released with inference code and a large, transparent evaluation dataset for reproducible benchmarking.
By Josef Liyanjun Chen
The paper evaluates how well typed decision models, which output probability distributions, actually provide reliable guarantees. By comparing open-weight Laya and Jev 1.13 across 190 evaluations, the study finds that point‑estimate thresholds frequently break promised accuracy, while a Beta lower‑bound rule and Learn‑then‑Test approach maintain guarantees more consistently. The work also explores calibration across languages, the impact of noise, data requirements for fine‑tuning, and the limitations of confidence in predicting outcomes or providing faithful explanations.
By Pedro Carvalho Brom
WorkFactory is a scalable framework that transforms professional intents into executable tasks by building an information support graph and using probe executions to detect and repair support failures. The authors generated 112,686 tasks across 612 occupations and fine‑tuned Qwen models (4B, 27B, 35B-A3B) to create the WorkPilot family, which outperforms base models on four professional‑work benchmarks. The study shows that planning and repair reduce task environment defects and improve downstream performance, with diminishing returns observed as model size increases.
By Ziyang Zeng, Tingyang Chen, Jia Deng, Qiyao Wang, Shuo Tang, Yike Yang, Tao Ren, Wayne Xin Zhao, Feng Chang, Jujie He, Chuan Hao, Ran Tao, Bryan Dai
The paper proposes a comprehensive data generation strategy to improve deep learning performance in classifying olive tree diseases. It combines traditional augmentation, CutMix, MixUp, and DCGAN-generated synthetic images to address data scarcity and class imbalance. Experiments with EfficientNetV2B0 show significant accuracy gains across three distribution scenarios, underscoring the importance of data diversity and balance.
By Hicham El Akhal, Abderrahim Waga, Aissa Ben Yahya, Abdelbaki El Belrhiti El Alaoui
Agent-Conditioned Routing (ACR) is a technique that enables a frozen sparse mixture-of-experts model to generate diverse reasoning paths from a single prompt without fine‑tuning or user‑assigned roles. By assigning persistent, untrained expert‑eligibility masks while keeping the model’s token router intact, ACR produces more problems with at least two distinct primary methods compared to temperature sampling on GLM 5.3 and GPT‑5.6 Sol. ACR adds minimal overhead (≈160 KiB) and maintains latency and VRAM usage within 3% of standard sampling.
By Ananth Kalyanasundaram, Ankush Kadu, Athira Gopal, Aswanth Krishnan
The paper introduces ReSeed, a lightweight post‑distillation technique that restores lost diversity in diffusion model distillation by adding a seed‑response bias adapter only to the first denoising step. By aligning the first step’s velocity with the teacher’s, ReSeed recovers most of the output diversity without altering later steps or adding target images. Experiments on SD3‑Medium, SD3.5‑Medium, FLUX.1‑dev, and Wan2.1‑T2V‑1.3B show a 48.3% increase in DINO diversity with negligible extra parameters and training data.
By Wenwu Tang, Dong Wang, Yun Cheng, Olga Saukh
The paper introduces HausaSafeLM, a safety preference dataset for Hausa comprising 10,000 prompts (3,000 human‑corrected and 7,000 machine‑translated) across six harm categories. Using Qwen2.5‑7B‑Instruct, the authors compare supervised fine‑tuning (SFT) and SFT followed by Direct Preference Optimization (DPO) on 500 harmful and 200 benign Hausa prompts, finding that SFT improves refusal fluency and precision—especially with larger, noisier data—while DPO sharpens precision at the cost of linguistic quality. No single approach simultaneously maximizes safety, helpfulness, and fluency, and the study shows that refusal‑rate metrics alone are unreliable indicators of safety in low‑resource languages.
By Valdini Douglace Lemofouet, Blessing Ngozi Uzor, Paula Chikaodinaka Anyanwu, Danielle Blanche Kapsa, Saminu Mohammad Aliyu, Seid Muhie Yimam, Shamsuddeen Hassan Muhammad
The study presents a multimodal machine‑learning framework that combines genomic sequences, epidemiological indicators, wastewater surveillance, and clinical rash images to detect measles and infer transmission pathways. In a three‑stage pipeline, the authors achieved high accuracy in classifying transmission linkage (AUC 0.924, F1 0.912), moderate success in predicting outbreak risk (AUC 0.707), and near‑perfect recall for aligning genomic and imaging embeddings (Recall@1 0.985). The results highlight the complementary strengths of genomic data for linkage prediction and wastewater data for outbreak‑risk forecasting.
By Jiada Li
The study evaluates the effectiveness of small, fine‑tuned open large language models (LLMs) for predicting bug severity in Eclipse bug reports, comparing them against traditional TF‑IDF classifiers and a fine‑tuned CodeBERT encoder. Using a balanced sample of 20,000 reports, eleven model configurations—including zero‑shot and QLoRA‑fine‑tuned Llama 3 8B and Qwen 2.5 7B—were tested on 2,000 held‑out reports, revealing that fine‑tuned LLMs achieve macro‑F1 scores around 0.52, comparable to CodeBERT, while zero‑shot LLMs lag behind TF‑IDF baselines. The analysis also highlights class‑specific performance differences and the impact of duplicate reports on model evaluation.
By Asset Kanatov, Talgar Bayan, Askar Boranbayev