Papers by Ivan P. Yamshchikov

8 papers
From Where Words Come: Efficient Regularization of Code Tokenizers Through Source Attribution (2026.acl-long)

Copied to clipboard

Challenge: Currently, subword tokenization is the most common approach for vocabulary building in large models.
Approach: They propose to regularize training and minimize overfitting by using source-attributed BPE . they find that undertrained tokens are prone to producing unused, unusable tokens .
Outcome: The proposed techniques reduce the number of under-trained tokens while maintaining the same inference procedure as with regular BPE.
Vygotsky Distance: Measure for Benchmark Task Similarity (2024.lrec-main)

Copied to clipboard

Challenge: GLUE, SuperGLUE and RussianSuperGLUE benchmarks are arbitrary sets of tasks that are not generalized.
Approach: They propose a theoretical instrument and an algorithm to calculate similarity between benchmark tasks . they use relative performance of the "students" on a given task to determine similarity .
Outcome: The proposed model reduces the number of evaluation tasks while maintaining high validation quality.
Teaching Old Tokenizers New Words: Efficient Tokenizer Adaptation for Pretrained Models (2026.findings-eacl)

Copied to clipboard

Challenge: Extending existing vocabulary is a widely used step in adapting pre-trained language models to new domains or languages.
Approach: They propose to extend a pre-trained tokenizer by continuing the BPE merge learning process on new data.
Outcome: The proposed method improves tokenization efficiency and improves model utilization.
ComicScene154: A Scene Dataset for Comic Analysis (2025.emnlp-main)

Copied to clipboard

Challenge: Comics offer compelling yet under-explored domain for computational narrative analysis . authors highlight potential of comics for narrative-driven, multimodal data analysis based on novel comics .
Approach: They propose a dataset of scene-level narrative arcs derived from comic books . they highlight their potential to inform broader research on multimodal storytelling .
Outcome: The dataset provides an initial benchmark that future studies can build upon.
Model in Distress: Sentiment Analysis on French Synthetic Social Media (2026.findings-acl)

Copied to clipboard

Challenge: Large companies and services handle a substantial volume of reviews and social media mentions.
Approach: They propose a generalizable pipeline for automated customer feedback analysis using backtranslation and fine-tuned models to generate 1.7 million tweets from a small seed corpus, complemented by synthetic reasoning traces.
Outcome: The proposed pipeline generates 1.7 million synthetic tweets from a small seed corpus, complemented by synthetic reasoning traces.
Smotrom tvoja på ander drogoj verden! Resurrecting Dead Pidgin with Generative Models: Russenorsk Case Study (2025.findings-acl)

Copied to clipboard

Challenge: Russenorsk is a pidgin language used in trade interactions between Russian and Norwegian speakers.
Approach: They propose to analyze Russenorsk's lexicon using modern large language models . they construct a structured dictionary of the language, grouped by synonyms and word origins .
Outcome: The proposed model can generate linguistic hypotheses about Russenorsk, the authors show . their model can be used to generate renderings of contemporary Russian and Norwegian texts .
Decomposing Textual Information For Style Transfer (D19-56)

Copied to clipboard

Challenge: Using a framework of style transfer for texts, we propose several empirical methods to assess information decomposition quality.
Approach: They propose to use latent representations to effectively decompose different aspects of textual information using a framework of style transfer for texts.
Outcome: The proposed methods show that higher quality representations correlate with higher performance in bilingual evaluation understudy (BLEU) between output and human-written reformulations.
Style Transfer for Texts: Retrain, Report Errors, Compare with Rewrites (D19-1)

Copied to clipboard

Challenge: Currently, standard methods for style transfer have several significant problems.
Approach: They propose to take BLEU between input and human-written reformulations into consideration for benchmarks.
Outcome: The proposed architectures outperform state-of-the-art in style transfer metric on human-written reformulations and take BLEU between input and output into consideration for benchmarks.

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations