Merge

Follow this topic Ask about this topic

Momentum

4 papers in the last four weeks, up 33% on the four weeks before. 0.0% of all new papers.

Jul 13Week of Sep 28

Latest papers 36

All topics

The Missing Coefficients: Bayesian Pairwise Merging for Model Personalization
Sep 30, 2026Yaling Shen, Tongtong Wu, Siyuan Yan +1Personalization Progress Reward Modeling
Monash University
Unmerge: Efficient Machine Unlearning via Task Arithmetic
Sep 30, 2026Haoran Tang, Andrew Tan, Rajiv KhannaMachine Unlearning Retraining
Department of Computer Science Purdue University
Action Conditioned Bisimulation For GUI Agent Memory
Sep 30, 2026Hongbo Zhang, Liuyang Song, Quanquan Li +3Graphical User Interface Agents Merge
Peking University · East China Normal University · University of Southern California
No Pain, More Gain: Iterative Merging for Effective Multi-Teacher On-Policy Distillation
Sep 28, 2026Seonghyeon Kim, Chaeyun Jang, Noah Lee +2Aware Heterogeneous Multi-Teacher Multimodal On-Policy Distillation Teacher
KAIST AI · Kakao
Better Together: Complementary Query Rewriting Under a Strong RAG Baseline
Sep 4, 2026Sara Shanian, Xiaoqin Yi, Pavlo Ruban +1Rewrite Large Language Model Baselines
ServiceNow
MergeOver: Post-Training Token Merging for Recursive Vision Transformers
Aug 13, 2026Junseo Kim, Uraz Odyurt, Amirreza YousefzadehSelf-Supervised Vision Transformers Vision Transformer
Computer Architecture for Embedded Systems, University of Twente Enschede, The Netherlands
Signature-Guided Capacity Occupancy for Dense Expert Merging
Aug 10, 2026Lingching Tung, Chi-Jui Kim, Beicheng Xu +2Continual Model Merging Merge
Peking University · Independent Researcher
SAFE-Merge: Data-Free Continual Model Merging with General Knowledge Preservation
Aug 2, 2026Zihuan Qiu, Zhiyang Liao, Chiyuan He +5Continual Model Merging Replay-Based Continual Learning
University of Electronic Science and Technology of China, Chengdu, China · Dalian University of Technology, Dalian, China
Mergeable Model-Side Aggregation States for Long-Context Language Models
Jul 29, 2026Dachuan Song, Junyu Yin, Zechen Hu +1Efficient Long-Context Inference Merge
Department of Electrical and Computer Engineering, George Mason University, Fairfax, VA, USA · Department of Computer Science, George Mason University, Fairfax, VA, USA
Entity Resolution in Practice: Lessons from a Self-Serve Pipeline
Jul 28, 2026Kaushik Pavani, Ganga Aluri, Pravin Jadhav +2Data Mining False-Negative Rate
Walmart
TwistedMerge: Certified Higher-Order Diagnostics and Abstention for Model Merging
Jul 23, 2026Ting Gong, Shitan XuContinual Model Merging Merge
Department of Mathematics, University of Washington, Box 354350, Seattle, WA 98195-4350, USA · Beijing International Center for Mathematical Research, Peking University, No. 5 Yiheyuan Road, Haidian District, Beijing, 100871, P.R. China
CT-Merging: Consensus Directions and Task-Level Scaling for LoRA Adapter Merging
Jul 18, 2026Keumseo Ryum, Joonhyuk KangContinual Model Merging Low-Rank Adaptation Adapters
KAIST
When Model Merging Rivals Joint Multi-Task Reinforcement Learning: A Task-Vector Geometry Analysis
Jul 17, 2026S. Aaron McClendonContinual Model Merging Merge
Aimpoint Digital Labs
Personalized Image Aesthetic Assessment via Preference-rich Sample Mining and Cohort Merging
Jul 17, 2026Zhichao Yang, Tianjiao Gu, Zhixianhe Zhang +3Aesthetics Merge
Xidian University Xi’an, China · Xidian University Xian, China
Do These Violent Delights Have Violent Ends? Measuring the Post-Merge Fate of Agentic Code
Jul 10, 2026Chunqiu Steven Xia, Courtney MillerPull Requests Coding Agents
University of Illinois Urbana-Champaign · Urbana, IL, USA · George Washington University +1
Cache Merging as a Convergent Replicated State for Multi-Agent Latent Reasoning
Jul 1, 2026Carlos Baquero, Luís BritoMerge Efficient Latent Reasoning
Faculdade de Engenharia (FEUP), Universidade do Porto (UP), Portugal · Escola Superior de Tecnologia e Gestão (ESTG), Politécnico de Viana do Castelo (IPVC), Portugal
Accelerating Merge with Motion Vector Difference via Filter Difference Analysis for VVenC
Jun 30, 2026Xinmin Feng, Shengyang Xu, Jianhua Chen +3Residual Vector Quantization Merge
University of Science and Technology of China · Hupan Laboratory
Fast Enough to Act: Spatio-Temporal Visual Token Merging for Low-Latency Robotic VLMs and VLAs
Jun 28, 2026Junzhou Chen, Jindong Wang, Gang ZhouDiffusion-Based Vision-Language-Actions Merge
Department of Computer Science, William & Mary, Williamsburg, VA 23188, USA · Department of Data Science, William & Mary, Williamsburg, VA 23188, USA
Post-Hoc Merging is Not Enough: Many-Shot Model Merging with Loss-Gap Balancing
Jun 15, 2026Kyungjin Im, Miru Kim, Chanin Eom +1Continual Model Merging Multi--Task Learning
Soongsil University, Republic of Korea · Department of Electrical and Computer Engineering, Sungkyunkwan University, Republic of Korea
Toward Instructions-as-Code: Understanding the Impact of Instruction Files on Agentic Pull Requests
Jun 11, 2026Ali Arabat, Mohammed SayaghPull Requests Coding Agents
École de Technologie Supérieure Montréal, Canada
GFFMERGE: Efficient Merging of Graph Neural Force Fields and Beyond
Jun 2, 2026Parth Verma, Parv P. Singh, Vipul Garg +3Graph Neural Networks Continual Model Merging
Department of Computer Science, Indian Institute of Technology Delhi, Hauz Khas, New Delhi, India · Yardi School of Artificial Intelligence, Indian Institute of Technology Delhi, Hauz Khas, New Delhi, India · Department of Civil Engineering, Indian Institute of Technology Delhi, Hauz Khas, New Delhi, India
ResMerge: Residual-based Spectral Merging of Large Language Models
Jun 1, 2026Yandu Sun, Zhiyan Hou, Haokai Ma +6Continual Model Merging Merge
1Southeast University, Nanjing, China · Institute of Automation, Chinese Academy of Sciences · University of Chinese Academy of Sciences +4
InfoMerge: Information-aware Token Compression for Efficient Video Large Language Models
Jun 1, 2026Xinxin Liu, Shiwei Gan, Xiao Liu +3Token Compression Video-Language Models
State Key Laboratory of Novel Software Technology, Nanjing University Nanjing 210023, China
MergeTok: Unified Continuous and Discrete Visual Tokenization via Token Merging
May 29, 2026Luyuan Zhang, Siyuan Li, Zedong Wang +7Visual Tokenizers Visual Generation
1Tsinghua University · 2Westlake University · 3Zhejiang University +3
Access Sets Matter: Budgeting Expert Reads for Scalable Weight-Space Model Merging
May 28, 2026Yuanyi Wang, Yanggan Gu, Su Lu +5Merge Inference Workloads
The Hong Kong Polytechnic University, PolyU · Hong Kong Polytechnic University Daya Bay Technology and Innovation Research Institute.
Extra-Merge: Tracing the Rank-1 Subspace of Model Merging in Language Model Pre-Training
May 26, 2026Wenjie Zhou, Bohan Wang, Hongtao Zhang +3Large Language Model Pretraining Merge
State Key Laboratory of AI Safety, Institute of Computing Technology, Chinese Academy of Sciences, China · University of Chinese Academy of Sciences, China · Alibaba Group, China +2
Merge-Bench: Resolve Merge Conflicts with Large Language Models
May 25, 2026Benedikt Schesch, Michael D. ErnstMerge
Amazon · University of Washington, USA
Unlocking the Potential of Continual Model Merging: An ODE Perspective
May 19, 2026Lihong Lin, Haidong KangContinual Model Merging Replay-Based Continual Learning
Northeastern University, Shenyang, China
MotionMERGE: A Multi-granular Framework for Human Motion Editing, Reasoning, Generation, and Explanation
May 18, 2026Bizhu Wu, Jinheng Xie, Wenting Chen +5Motion Merge
Computer Vision Institute, School of Computer Science and Software Engineering, Shenzhen University, Shenzhen 518060, China · Guangdong Key Laboratory of Intelligent Information Processing, Shenzhen University, Shenzhen 518060, China · School of Computer Science, University of Nottingham Ningbo China +4
Rover: Context-aware Conflict Resolution with LLM
May 17, 2026Qingyu Zhang, Junzhe Li, Jiayi Lin +2Merge Autonomous Surface Vehicles
The University of Hong Kong, China · Wuhan University, China
Scaling Retrieval-Augmented Reasoning with Parallel Search and Explicit Merging
May 13, 2026Jiabei Liu, Wenyu Mao, Junfei Tan +4Retrievers Search Agents
University of Science and Technology of China · WeChat Technical Architecture Department, Tecent Inc.
Merging Methods for Multilingual Knowledge Editing for Large Language Models: An Empirical Odyssey
May 13, 2026Kunil Lee, Ki-Young Shin, Jong-Hyeok Lee +1Multimodal Knowledge Editing Multilingual Language Models
Department of Computer Science and Engineering, POSTECH · Designovel Co., Ltd. · LLSOLLU +1
ORBIT: Preserving Foundational Language Capabilities in GenRetrieval via Origin-Regulated Merging
May 12, 2026Neha Verma, Nikhil Mehta, Shao-Chuan Wang +7Large Language Model Fine-Tuning Catastrophic Forgetting
Johns Hopkins University · Google DeepMind · Google
PhySPRING: Structure-Preserving Reduction of Physics-Informed Twins via GNN
May 8, 2026Yixiong Jing, Xingyuan Chen, Guangming Wang +3Digital Twins Physics Simulation
University of Cambridge · Shanghai Jiao Tong University
Train Separately, Merge Together: Modular Post-Training with Mixture-of-Experts
Apr 20, 2026Jacob Morrison, Sanjay Adhikesaven, Akshita Bhagia +3Mixture-Of-Experts Post-Training
University of Washington · Allen Institute for AI · University of California, Berkeley
K-Merge: Online Continual Merging of Adapters for On-device Large Language Models
Oct 15, 2025Donald Shenaj, Ondrej Bohdal, Taha Ceritli +3Low-Rank Adaptation Adapters Large Language Model Adaptation
Samsung R&D Institute UK · University of Pisa · University of Padova

Sep 30, 2026, cs.LGJ/K move · Enter open · S save

The Missing Coefficients: Bayesian Pairwise Merging for Model Personalization

Yaling Shen, Tongtong Wu, Siyuan Yan, Gholamreza Haffari

Monash University

Personalization Progress Reward Modeling Merge Latent Variable Missing Data

Open paper AskPDF arXiv

How can we personalize a shared expert library from a user's pairwise choices? Prior work can realize different reward trade-offs by merging reward-specialized experts, given a vector of trade-off weights. In practice, users can more naturally choose between outputs than specify numerical weights. The challenge is therefore to turn these choices into the coefficients required for merging, while accounting for ambiguity when feedback is limited. Our key idea is to treat the unknown reward weights as latent variables: infer a posterior over them from pairwise choices and reward-score differences, and use its mean directly as the merge coefficients. We instantiate this idea as Bayesian Pairwise Merging (BPM), whose posterior also characterizes which reward trade-offs remain plausible given the feedback. We evaluate BPM on radiology summarization, image captioning, and story generation, spanning text-to-text and image-to-text generation. With 100 feedback per simulated persona, BPM achieves macro decided win rates of 91.7%, 77.1%, and 64.3% against uniform merge. For six pairs of simulated personas, each prefers the model fitted to its own feedback, a pattern also observed in a human proof-of-concept. In simulations under BPM's model and prior, its nominal 90% intervals for temperature-scaled reward weights achieve task-averaged marginal coverage of 88.9% and 89.2% with only 10 and 25 comparisons, respectively. BPM thus enables personalization from pairwise feedback without per-user policy training, while characterizing the coefficient ambiguity left by limited feedback.

Momentum

Related topics

Latest papers 36