LoRA, PEFT, instruction tuning and domain adaptation — adapting a pretrained model without paying to train one.
arXiv:2510.17198v2 Announce Type: replace Abstract: Riverbank erosion is a serious environmental problem in Bangladesh, causing land loss, damage to infrastructure, and displacement of local communit...
arXiv:2608.28161v1 Announce Type: cross Abstract: Mango variety identification in Bangladesh is challenging because closely related cultivars can have similar visual characteristics and images are of...
arXiv:2606.21359v2 Announce Type: replace Abstract: Large language models (LLMs) are increasingly used to communicate and explain scientific concepts, yet their tendency to hallucinate poses signific...
The paper introduces Qwen-GuidePlay-2B, a 2B‑parameter language model fine‑tuned for dialogue‑game interaction. The training pipeline consists of three stages: supervised fine‑tuning on successful game trajectories, weighted turn‑level fine‑tuning, and teacher‑guided fine‑tuning that corrects formatting and evaluates examples. The resulting model achieves a clemscore of 57.12 and a statscore of 42.68 on Playpen’s public validation, ranking second in the official challenge and demonstrating that careful curation can outperform more aggressive procedural methods.
The paper introduces MSCGC-KAN, a new EEG emotion recognition approach that builds on a pre‑trained CBraMod backbone. It incorporates a structured task head featuring multi‑scale causal graph convolution and Kolmogorov–Arnold feature mapping to better capture multi‑scale emotional dynamics, inter‑channel connectivity, and nonlinear discriminative patterns. Experiments on FACED and SEED‑VII show significant performance gains over a linear baseline, achieving balanced accuracies of 60.66% and 33.27% respectively.
The paper addresses the mismatch between learner and client data distributions in federated learning, noting that traditional client selection methods often ignore this misalignment. It introduces a dynamic, influence-aware client selection framework that uses a small proxy dataset to estimate each client's utility for the learner’s objective, prioritizing informative sources while mitigating noise and heterogeneity. Experiments on CIFAR-10 with heterogeneous partitions show the proposed method outperforms static and dynamic baselines, achieving faster convergence and higher accuracy.
The paper presents a post‑training recipe for small dialogue‑game agents that involves three steps: acquiring broad game participation via supervised fine‑tuning, repairing specific local failures with turn‑local preference pairs, and preserving general capabilities. Applied to the LM Playschool Challenge, the method raises the public clemscore from 10.67 to 38.92 and the closed in‑domain score from 13.41 to 41.17 while keeping overall static performance nearly unchanged. The gains are mainly within the targeted game family, with limited improvement on out‑of‑domain clemscore.
RegCL is a non‑replay continual learning framework that adapts the Segment Anything Model (SAM) for visual grounding across evolving multi‑sensorial media domains. It consolidates domain‑specific segmentation knowledge into a single lightweight SAM adapter by incrementally merging LoRA‑style AugModules and preserving compact historical feature statistics. Experiments on five heterogeneous datasets demonstrate that RegCL retains performance while adapting to new domains, outperforming other non‑replay continual learning and merging baselines.
The paper presents an explainable diabetic retinopathy classification framework that leverages vision foundation models—DINOv2, CLIP, and Vision Transformer—combined with various transfer learning techniques such as full fine‑tuning, linear probing, and Low‑Rank Adaptation (LoRA). Using the ODIR dataset for internal validation and the APTOS dataset for external testing, DINOv2‑LoRA achieved the best internal AUROC (0.758) while DINOv2 and ViT full fine‑tuning reached the highest external AUROC (0.920). Explainability was assessed with Grad‑CAM and HiResCAM against expert‑annotated lesion masks from IDRiD, using Dice, IoU, and Pointing Game metrics, confirming that model attention aligns with clinically relevant retinal lesions.
The paper introduces AIM, a two‑stage approach for unlearning identity‑specific information from multimodal large language models (MLLMs) when retain images are not available at deletion time. AIM first anchors an identity‑forgetting target using a universal visual prompt, then aligns the vision encoder to this target under a Fisher‑based constraint. Experiments demonstrate that AIM effectively removes identity knowledge while preserving other visual perception capabilities and prior knowledge.
The paper introduces Short‑Films 20K (SF20K), a large publicly available movie dataset comprising 20,143 amateur films totaling 3,582 hours, with an average length of 12 minutes per film. Accompanying the dataset is SF20K‑Test, a manual open‑ended question‑answering benchmark featuring 95 movies and 979 question‑answer pairs. Analysis of the benchmark shows limited data leakage, highlights the necessity of long‑term reasoning, and demonstrates that instruction tuning on the large‑scale dataset significantly boosts vision‑language model performance.
The paper introduces DARTS, a method for tuning decoder representations during model merging. It addresses representation bias in autoregressive decoders by using an entropy‑weighted L1 loss and a per‑position additive bias to correct errors that accumulate across token positions. Experiments on code generation, mathematical reasoning, and instruction following with Llama‑2‑7B show that DARTS improves performance over standard surgery while adding only 0.1% extra parameters.
The paper introduces SPAR‑Bench, a set of eight probes designed to test whether medical vision models can reason about anatomy in abdominal CT scans. Experiments across five architectures and three foundation models—both frozen and fine‑tuned—show that while models can recall canonical organ locations, they fail to perform relational reasoning or spatial comparisons within a patient, even under zero‑shot transfer. The study also demonstrates that pooled probing underestimates a model’s relational capabilities and that open‑weight multimodal large language models perform poorly on these tasks.
The paper introduces Learned Relay Representations (Relay), a technique for Masked Diffusion Models (MDMs) that preserves and forwards internal latent information across denoising steps via a differentiable per-token channel trained with truncated backpropagation through time. Relay enables MDMs to be forward‑thinking, avoiding costly recomputation of internal representations and improving efficiency. Applied to Fast‑dLLM v2, Relay outperforms standard supervised fine‑tuning on coding tasks and reduces inference latency by up to 32%, demonstrating a clear performance‑latency advantage for diffusion language models.
Aero Hand Open is a tendon‑driven, anthropomorphic hand designed for affordable, simulation‑ready dexterous manipulation. The release includes a realistic simulation model of the cable transmission, an identified actuation map that links motor commands to joint motions (including thumb coupling), and a reinforcement‑learning package that trains policies entirely in simulation. These components enable policies to run on the physical hand without fine‑tuning or state estimation.
The paper introduces ‘One Model for All’, a universal pre‑training framework that tackles EEG‑based emotion recognition across diverse datasets and paradigms. It decouples learning into a univariate self‑supervised contrastive pre‑training stage using a Unified Channel Schema, followed by a multivariate fine‑tuning stage that employs an Adaptive Resampling Transformer and a Graph Attention Network to model spatio‑temporal dependencies. Experiments demonstrate state‑of‑the‑art performance on within‑subject benchmarks (SEED 99.27%, DEAP 93.69%, DREAMER 93.93%) and superior cross‑dataset transfer, with ablation studies highlighting the critical role of the GAT module.
SemEnrich introduces a self‑supervised method to enrich radiology reports by clustering sentences semantically and adding positive or neutral observations from different clusters. The enriched data consistently improves supervised fine‑tuning across multiple vision‑language metrics, with gains ranging from 3% to 7.5% on COMET, Bert, Sentence Bleu, CheXbert‑F1, and RadGraph‑F1. The authors also demonstrate that incorporating cluster information into the reward design for GRPO training yields additional performance boosts.
D-TAIA is a framework that adapts large language models for multi‑task predictive process monitoring, jointly predicting the next activity and remaining time of ongoing cases. It uses domain‑aware triplet loss pre‑training, FAISS‑based nearest‑neighbor retrieval for time estimation, and a TAIA inference strategy to preserve sequential reasoning while fine‑tuning a 10 M‑parameter backbone. Across four real‑world event logs, D‑TAIA achieves state‑of‑the‑art or competitive results compared to a fine‑tuned LLM and a recurrent neural network baseline, with ablation studies showing the effectiveness of NLP and computer‑vision techniques for this domain.
The paper introduces VLAct, a Vision‑Language‑Action model that focuses on representation‑centric continued pre‑training rather than merely scaling robot data. VLAct is trained on diverse, multi‑embodiment robot data and preserves a broad VLM prior while encouraging shared action semantics across embodiments. Experiments across simulation, real‑world, and unseen‑embodiment settings show that VLAct consistently outperforms existing industrial VLA systems, achieving high success rates with only a modest compute budget and open‑source data.
arXiv:2608.28237v1 Announce Type: new Abstract: Predictive Process Monitoring (PPM) models are increasingly deployed in dynamic environments where concept drift causes the underlying process distribu...