Nov 2026· Proceedings of the 34th ACM International Conference on Multimedia· 0 citations· 5 references
TL;DR
This work identifies a key bottleneck termed attention dilution and forms the extraction of task-relevant semantic focus as a variational inference problem, and proposes Ruminative Contrastive Attention Calibration (RumiA), a training-free inference method that achieves competitive improvement over strong baselines and reduces the computational cost.
Abstract
While Multimodal Large Language Models excel at general visual question answering, their performance degrades significantly on tasks that depend on fine-grained details. This limitation makes it challenging for models to reliably handle tasks such as counting, document understanding, and complex reasoning. A recent strategy to enhance visual details is to perform multi-round inference, using the model's attention to crop a single, rectangular region of interest. This approach is computationally inefficient and is less effective for tasks that require analyzing multiple, arbitrarily shaped regions. To overcome these flaws, we identify a key bottleneck termed attention dilution and formulate the extraction of task-relevant semantic focus as a variational inference problem. Derived from the optimal solution under this formulation, we propose Ruminative Contrastive Attention Calibration (RumiA), a training-free inference method. RumiA computes the Conditional Pointwise Mutual Information of visual tokens via a Layer-Wise Attention Contrast, acting as a semantic unsharp masking filter that helps suppress background priors while maintaining the ability to attend to multiple regions. Capitalizing on the cross-scale stability of this distribution, RumiA leverages an efficient, smaller MLLM to perform a soft key-value calibration on a larger target model during a single inference pass. We evaluated our method on three MLLM families across ten benchmarks, focusing on fine-grained and complex reasoning. Our results show that RumiA achieves competitive improvement over strong baselines and reduces the computational cost.
Supporting data, adapters, predictions and code for the article *Low-Cost LoRA Fine-Tuning of Small Language Models for Multi-Step Arithmetic Reasoning* by Jake O'Grady, Asena Isik Gürhan, Chee Fong Ting and Effirul Ramlan (University of Galway). We generated 20,000 GSM8K-derived arithmetic problems with step-by-step s...
O'Grady, Jake, Gürhan, Asena Isik, Chee, Fong Ting et al.· Zenodo (CERN European Organi...· 465 citations
The results are packaged in the Greenfield Startup Model (GSM), which explains the priority of startups to release the product as quickly as possible, and the need to shorten time-to-market, by speeding up the development through low-precision engineering activities.
Carmine Giardino, Nicolò Paternoster, M. Unterkalmsteiner et al.· IEEE Transactions on Softwar...· 178 citations· ⚡14
Software startup companies develop innovative, software-intensive products within limited timeframes and with few resources, searching for sustainable and scalable business models.
M. Unterkalmsteiner, P. Abrahamsson, Xiaofeng Wang et al.· e-Informatica Software Engin...· 157 citations· ⚡17
This study conducts a case survey study based on the secondary data of the major pivots happened in 49 software startups, and demonstrates that customer need pivot is the most common among all pivot types.
Sohaib Shahid Bajwa, Xiaofeng Wang, Anh Nguyen-Duc et al.· Empirical Software Engineeri...· 127 citations· ⚡15
The comparison of adopter and non-adopter sample reveals three potential adoption inhibitor, security, data privacy, and portability, which underlines the importance of the technical and security perspectives for research investigating the adoption of technology.
Nattakarn Phaphoom, Xiaofeng Wang, S. Samuel et al.· Journal of Systems and Softw...· 111 citations· ⚡8
Objective There is dearth of information on the timeliness of antenatal care (ANC) uptake. This study aimed to determine the timely ANC uptake by a medically trained provider (MTP) as per the World Health Organization (WHO) recommendations and the country guideline. Methods Cross-sectional survey was done with 2,731 wo...
B. Sarker, Musfikur Rahman, T. Rahman et al.· PLoS ONE· 89 citations· ⚡7
Related blog posts
Microsoft Research Blog· microsoft.comNov 16, 2026
AI coding agents are increasingly help developers fix bugs and build features largely on their own, but every interaction with their models consumes tokens. In each session, an agent sends instructions, tool descriptions, and session history to its model. Most efforts to reduce this cost focus on individual tasks. Yet developers return to the same […] The post TokenEconomizer: Learning from Experience to Make AI Coding Agents More Efficient appeared first on Microsoft Research.
MIT scientists developed a training method to generate models that simulate the effects of human activities on climate much more quickly than other approaches.
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.