Nov 2026· Proceedings of the 35th ACM International Conference on Information and Knowledge Management· 0 citations· 35 references
Privacy-Preserving Technologies in Data
Abstract
Federated class-incremental learning (FCIL) aims to learn emerging classes across decentralized clients while preserving previously acquired knowledge. Existing FCIL methods are often designed or evaluated under relatively synchronous new-class exposure settings, while in realistic federated systems, a new class may initially appear on only a few clients with limited samples. This creates a local-global propagation mismatch: scarce new-class samples make local updates unstable, whereas majority-dominated aggregation dilutes these updates and slows the global propagation of new-class knowledge. To address this problem, we propose Hierarchical Contribution-Aware Aggregation (FedHCA), a novel framework for FCIL under asynchronous new-class exposure. FedHCA stabilizes local new-class update formation via a prototype-guided local learning strategy, and further promotes global new-class propagation through coarse-to-fine hierarchical aggregation. Specifically, the server first distinguishes new-class-exposed clients from old-class-only clients, then groups exposed clients according to their semantic and update-direction consistency, and finally performs contribution-weighted group aggregation based on update magnitudes. This design amplifies informative new-class updates while maintaining old-class stability. Experiments on CIFAR-100, CIFAR-10, and SVHN show that FedHCA achieves the best overall accuracy and a better balance between old-class retention and new-class adaptation compared with representative baselines. Further analyses confirm its propagation efficiency, component effectiveness, and robustness under asynchronous exposure.
The results indicate that software engineering work practices are chosen opportunistically, adapted and configured to provide value under the constrains imposed by the startup context.
Nicolò Paternoster, Carmine Giardino, M. Unterkalmsteiner et al.· Information and Software Tec...· 394 citations· ⚡54
This state-of-practice investigation was performed using a literature review followed by a multiple-case study approach and presents how inconsistency between managerial strategies and execution can lead to failure by means of a behavioral framework.
Carmine Giardino, Xiaofeng Wang, P. Abrahamsson· International Conference on...· 175 citations· ⚡19
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
It is found that roles of MVPs in startups were not fully aware by entrepreneurs, and entrepreneurs should consider a systematic approach to fully explore the value of MVP, as a multiple facet product (MFP).
Anh Nguyen-Duc, P. Abrahamsson· International Conference on...· 93 citations· ⚡9
It is found that what perceived as biggest challenges by software startups do vary across different life cycle stages, even though its significance decreases when the learning focuses of the startups move from problem to solution and their products mature.
Xiaofeng Wang, Henry Edison, Sohaib Shahid Bajwa et al.· International Conference on...· 62 citations· ⚡6
A comprehensive overview of how enhanced sampling methods are reshaping the field, with a particular focus on the data-driven construction of collective variables, is provided.
Kai Zhu, Enrico Trizio, Jintu Zhang et al.· Chemical Reviews· 58 citations
Related blog posts
MIT News · Artificial Intelligence· news.mit.eduNov 17, 2026
By studying the computational foundations of language learning, Associate Professor Jacob Andreas builds intelligent systems that can communicate effectively with humans.
MIT News · Artificial Intelligence· news.mit.eduNov 17, 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 News · Artificial Intelligence· news.mit.eduNov 9, 2026