5 papers
LLMs are General Asynchronous Agents
George Yakushev, Denis Mazur, Vladimir Bartenev +4
Modern LLMs are increasingly capable as autonomous agents, but they follow sequential interaction cycles: read, think, reply or call tools, repeat. Many real-world use cases are no…
On Efficient Scaling of GNNs via IO-Aware Layers Implementations
Daria Fomina, Daniil Krasylnikov, Alexey Boykov +3
Graph Neural Networks (GNNs) are bottlenecked by sparse, irregular memory access. Popular frameworks such as DGL and PyTorch Geometric support general message passing, but complex…
KV Cache Offloading for Context-Intensive Tasks
Andrey Bocharnikov, Ivan Ermakov, Denis Kuznedelev +2
With the growing demand for long-context LLMs across a wide range of applications, the key-value (KV) cache has become a critical bottleneck for both latency and memory usage. Rece…
Asynchronous Reasoning: Training-Free Interactive Thinking LLMs
George Yakushev, Nataliia Babina, Masoud Vahid Dastgerdi +4
Many state-of-the-art LLMs are trained to think before giving their answer. Reasoning can greatly improve language model capabilities, but it also makes them less interactive: give…
Predicting Performance of Heterogeneous AI Systems with Discrete-Event Simulations
Vyacheslav Zhdanovskiy, Lev Teplyakov, Anton Grigoryev
In recent years, artificial intelligence (AI) technologies have found industrial applications in various fields. AI systems typically possess complex software and heterogeneous CPU…