Papers by Aydar Bulatov
Cramming 1568 Tokens into a Single Vector and Back Again: Exploring the Limits of Embedding Space Capacity (2025.acl-long)
Copied to clipboard
| Challenge: | Recent work addresses problem of compression of tokens into shorter sequence of real-valued vectors . attainable lossless compression ratio is typically not higher than x10 . |
| Approach: | They propose to compress a sequence of tokens into a shorter sequence of real-valued vectors to be used as inputs instead of token embeddings or key-value cache. |
| Outcome: | The proposed algorithms reduce the amount of compute in existing language models rather than minimizing number of bits needed to store text. |
Better Together: Enhancing Generative Knowledge Graph Completion with Language Models and Neighborhood Information (2023.findings-emnlp)
Copied to clipboard
| Challenge: | Knowledge graph completion (KGC) methods are computationally intensive and impractical for large-scale KGs. |
| Approach: | They propose to include node neighborhoods as additional information to improve KGC methods based on language models. |
| Outcome: | The proposed method outperforms KGT5 and conventional methods on inductive and transductive Wikidata subsets and shows its importance. |
Wikontic: Constructing Wikidata-Aligned, Ontology-Aware Knowledge Graphs with Large Language Models (2026.eacl-long)
Copied to clipboard
| Challenge: | Knowledge graphs provide structured, verifiable grounding for large language models . current LLMs use KGs as auxiliary structures for text retrieval . |
| Approach: | They propose a pipeline that constructs KGs from open-domain texts using triplets and qualifiers. |
| Outcome: | The proposed pipeline outperforms existing methods in retrieval-augmented generation. |
Beyond Memorization: Extending Reasoning Depth with Recurrence, Memory and Test-Time Compute Scaling (2026.findings-acl)
Copied to clipboard
Ivan Rodkin, Daniil Orel, Konstantin Smirnov, Arman Bolatov, Bilal Elbouardi, Besher Hassan, Yuri Kuratov, Aydar Bulatov, Preslav Nakov, Timothy Baldwin, Artem Shelmanov, Mikhail Burtsev
| Challenge: | Reasoning is a core capability of large language models, yet how multi-step reasoning is learned and executed remains unclear. |
| Approach: | They evaluate how large language models learn multi-step reasoning without memorization . they find that most neural architectures trained from scratch can learn rule inference . |
| Outcome: | The proposed framework fails to solve a natural-language proxy task with high accuracy. |