5 papers
Evidence for Shared Routing Geometry and Dynamics in Sparse Mixture-of-Experts
Kirill Labzin, Stepan Kulibaba, Artem Dzhalilov +1
Sparse mixture-of-experts (MoE) models use an independently parameterized router at each sparse layer to select experts for every token. Prior work has shown that routing decisions…
SDG-MoE: Signed Debate Graph Mixture-of-Experts
Stepan Kulibaba, Kirill Labzin, Artem Dzhalilov +4
Sparse MoE models achieve a good balance between capacity and compute by routing each token to a small subset of experts. However, in most MoE architectures, once a token is routed…
AdLoCo: adaptive batching significantly improves communications efficiency and convergence for Large Language Models
Nikolay Kutuzov, Makar Baderko, Stepan Kulibaba +4
Scaling distributed training of Large Language Models (LLMs) requires not only algorithmic advances but also efficient utilization of heterogeneous hardware resources. While existi…
KompeteAI: Accelerated Autonomous Multi-Agent System for End-to-End Pipeline Generation for Machine Learning Problems
Stepan Kulibaba, Artem Dzhalilov, Roman Pakhomov +3
Recent Large Language Model (LLM)-based AutoML systems demonstrate impressive capabilities but face significant limitations such as constrained exploration strategies and a severe…
Exploring Applications of State Space Models and Advanced Training Techniques in Sequential Recommendations: A Comparative Study on Efficiency and Performance
Mark Obozov, Makar Baderko, Stepan Kulibaba +2
Recommender systems aim to estimate the dynamically changing user preferences and sequential dependencies between historical user behaviour and metadata. Although transformer-based…