Results 21 to 30 of about 9,649,421 (248)

Hybrid Spiking LoRA: Asymmetric Bit-Width Design for Language Model Adaptation With Theoretical Neuromorphic Efficiency Potential—A Case Study on Korean NLU

IEEE Access
Parameter-efficient fine-tuning adapts pre-trained language models to downstream tasks with reduced training cost, but widely used methods such as Low-Rank Adaptation (LoRA) still rely on conventional multiply-accumulate computation in the adapter path ...
Jae-Hwan Kim, Dae-Yeol Kim, Chae-Bong Sohn   +2 more
doaj   +1 more source

KDM7A and KDM1A inhibition suppresses tumour promoting pathways in prostate cancer

Molecular Oncology, EarlyView.
Treatment resistance is a major challenge for patients with advanced prostate cancer. This study examined an alternative approach to target the major prostate cancer‐promoting pathway by targeting epigenetic factors, whose levels are higher in tumours.
Jennie N Jeyapalan, Veronika M Metzler, Simone de Brot, Corinne L Woodcock, Anna E Harris, Jennifer Lothion‐Roy, Emeli M Nilsson, Atara Ntekim, Michael S Toss, Jenny L Persson, Francesca Khani, Brian D Robinson, Lorraine J Gudas, Emad Rakha, David M Heery, Catrin S Rutland, Nigel P Mongan   +16 more
wiley   +1 more source

Lottery Rank-Pruning Adaptation Parameter Efficient Fine-Tuning

Mathematics
Recent studies on parameter-efficient fine-tuning (PEFT) have introduced effective and efficient methods for fine-tuning large language models (LLMs) on downstream tasks using fewer parameters than required by full fine-tuning.
Juhyeong Kim, Gyunyeop Kim, Sangwoo Kang
doaj   +1 more source

MTL-LoRA: Low-Rank Adaptation for Multi-Task Learning

Proceedings of the AAAI Conference on Artificial Intelligence
Parameter-efficient fine-tuning (PEFT) has been widely employed for domain adaptation, with LoRA being one of the most prominent methods due to its simplicity and effectiveness. However, in multi-task learning (MTL) scenarios, LoRA tends to obscure the distinction between tasks by projecting sparse high-dimensional features from different tasks into ...
Yaming Yang 0001, Dilxat Muhtar, Yelong Shen, Yuefeng Zhan, Jianfeng Liu, Yujing Wang 0002, Hao Sun 0015, Denvy Deng, Feng Sun 0008, Qi Zhang 0066, Weizhu Chen, Yunhai Tong   +11 more
openaire   +4 more sources

Is an Apple an Orange? A Large Language Model Benchmark for Candidate Term Extraction and Subclass Decisions Against Upper Ontologies in Engineering and Materials Science

Advanced Engineering Materials, EarlyView.
Building machine‐readable vocabularies for materials science is slow, expert‐driven work. This study benchmarks 13 large language models on two of its first steps: finding candidate terms in engineering articles and deciding where they belong in a class hierarchy.
Thomas Bjarsch, Maximilian Holland, Klaus Drechsler, Johannes Schilp   +3 more
wiley   +1 more source

Efficient Low-Rank Adaptation for Sparse Large Language Model

Tsinghua Science and Technology
Existing Low-Rank Adaptation (LoRA) methods face challenges on sparse Large Language Models (LLMs) due to the inability to maintain sparsity. Recent works introduce methods that maintain sparsity by augmenting LoRA techniques with additional masking ...
Yuxuan Hu, Tian Tian, Xiaodong Chen, Zhe Zhao, Tao Tao, Weifang Zhang, Yuanfeng Li, Yuhang Liang, Cuiping Li, Hong Chen, Jing Zhang   +10 more
doaj   +1 more source

Grounding Large Language Models for Robot Task Planning Using Closed‐Loop State Feedback

Advanced Robotics Research, EarlyView.
BrainBody‐Large Language Model (LLM) introduces a hierarchical, feedback‐driven planning framework where two LLMs coordinate high‐level reasoning and low‐level control for robotic tasks. By grounding decisions in real‐time state feedback, it reduces hallucinations and improves task reliability.
Vineet Bhat, Ali Umut Kaypak, Prashanth Krishnamurthy, Ramesh Karri, Farshad Khorrami   +4 more
wiley   +1 more source

Performance and Computational Cost of Full and Parameter-Efficient Fine Tuning for Arabic Sentiment Classification Across Training Set Sizes

Computation
Pre-trained language models are typically adapted to downstream tasks via full fine tuning. However, this entails substantial computational and memory overhead.
Teif Aldaajani, Morooj Alqurashi, Sarah Aljuaid, Maha Jarallah Althobaiti   +3 more
doaj   +1 more source

Stable-LoRA: Stabilizing Feature Learning of Low-Rank Adaptation

CoRR
Low-Rank Adaptation (LoRA) is a widely adopted parameter-efficient method for fine-tuning Large Langauge Models. It updates the weight matrix as $W=W_0+sBA$, where $W_0$ is the original frozen weight, $s$ is a scaling factor and $A$,$B$ are trainable low-rank matrices.
Yize Wu, Ke Gao, Ling Li, Yanjun Wu
openaire   +3 more sources

Sustainable Materials Design With Multi‐Modal Artificial Intelligence

Advanced Science, EarlyView.
Critical mineral scarcity, high embodied carbon, and persistent pollution from materials processing intensify the need for sustainable materials design. This review frames the problem as multi‐objective optimization under heterogeneous, high‐dimensional evidence and highlights multi‐modal AI as an enabling pathway.
Tianyi Xu, Tianshuo Wei, Yan Ge, Bo Peng, Yue Li, Maolin Wang, Peng Wen, Chao Yang, Ye Wei   +8 more
wiley   +1 more source