Dmitry Ustalov

Dmitry Ustalov, Nikita Pavlichenko, Sergey Koshelev et al.

In this paper, we present Toloka Visual Question Answering, a new crowdsourced dataset allowing comparing performance of machine learning systems against human level of expertise in the grounding visual question answering task. In this task, given an image and a textual question, one has to draw the bounding box around the object correctly responding to that question. Every image-question pair contains the response, with only one correct response per image. Our dataset contains 45,199 pairs of images and questions in English, provided with ground truth bounding boxes, split into train and two test subsets. Besides describing the dataset and releasing it under a CC BY license, we conducted a series of experiments on open source zero-shot baseline models and organized a multi-phase competition at WSDM Cup that attracted 48 participants worldwide. However, by the time of paper submission, no machine learning model outperformed the non-expert crowdsourcing baseline according to the intersection over union evaluation score.

Nikita Pavlichenko, Dmitry Ustalov

Recent progress in generative models, especially in text-guided diffusion models, has enabled the production of aesthetically-pleasing imagery resembling the works of professional human artists. However, one has to carefully compose the textual description, called the prompt, and augment it with a set of clarifying keywords. Since aesthetics are challenging to evaluate computationally, human feedback is needed to determine the optimal prompt formulation and keyword combination. In this paper, we present a human-in-the-loop approach to learning the most useful combination of prompt keywords using a genetic algorithm. We also show how such an approach can improve the aesthetic appeal of images depicting the same descriptions.

Dmitry Ustalov

The rapid advancement of natural language processing (NLP) technologies, such as instruction-tuned large language models (LLMs), urges the development of modern evaluation protocols with human and machine feedback. We introduce Evalica, an open-source toolkit that facilitates the creation of reliable and reproducible model leaderboards. This paper presents its design, evaluates its performance, and demonstrates its usability through its Web interface, command-line interface, and Python API.

Georgii Levtsov, Dmitry Ustalov

With the advent of highly capable instruction-tuned neural language models, benchmarking in natural language processing (NLP) is increasingly shifting towards pairwise comparison leaderboards, such as LMSYS Arena, from traditional global pointwise scores (e.g., GLUE, BIG-bench, SWE-bench). This paper empirically investigates the strengths and weaknesses of both global scores and pairwise comparisons to aid decision-making in selecting appropriate model evaluation strategies. Through computational experiments on synthetic and real-world datasets using standard global metrics and the popular Bradley-Terry model for pairwise comparisons, we found that while global scores provide more reliable overall rankings, they can underestimate strong models with rare, significant errors or low confidence. Conversely, pairwise comparisons are particularly effective for identifying strong contenders among models with lower global scores, especially where quality metrics are hard to define (e.g., text generation), though they require more comparisons to converge if ties are frequent. Our code and data are available at https://github.com/HSPyroblast/srw-ranking under a permissive license.

Dmitry Ustalov, Egor Bogomolov, Alexander Bezzubov et al.

The rapid advancement of workflows and methods for software engineering using AI emphasizes the need for a systematic evaluation and analysis of their ability to leverage information from entire projects, particularly in large code bases. In this challenge on optimization of context collection for code completion, organized by JetBrains in collaboration with Mistral AI as part of the ASE 2025 conference, participants developed efficient mechanisms for collecting context from source code repositories to improve fill-in-the-middle code completions for Python and Kotlin. We constructed a large dataset of real-world code in these two programming languages using permissively licensed open-source projects. The submissions were evaluated based on their ability to maximize completion quality for multiple state-of-the-art neural models using the chrF metric. During the public phase of the competition, nineteen teams submitted solutions to the Python track and eight teams submitted solutions to the Kotlin track. In the private phase, six teams competed, of which five submitted papers to the workshop.

Dmitry Ustalov, Nikita Pavlichenko, Boris Tseitlin

This paper presents Crowd-Kit, a general-purpose computational quality control toolkit for crowdsourcing. Crowd-Kit provides efficient and convenient implementations of popular quality control algorithms in Python, including methods for truth inference, deep learning from crowds, and data quality estimation. Our toolkit supports multiple modalities of answers and provides dataset loaders and example notebooks for faster prototyping. We extensively evaluated our toolkit on several datasets of different natures, enabling benchmarking computational quality control methods in a uniform, systematic, and reproducible way using the same codebase. We release our code and data under the Apache License 2.0 at https://github.com/Toloka/crowd-kit.

Nikita Pavlichenko, Iurii Nazarov, Ivan Dolgov et al.

We present the Mellum models family, open-weight code completion models designed for interactive use in JetBrains IDEs. Mellums have 4B parameters, adopt a Llama-style architecture, and are pre-trained on ~4T tokens of permissively licensed, multi-language code. Our studies show that (i) careful data curation and staged training significantly improve the model's quality, (ii) editor-critical capabilities such as context packing are necessary for high-quality suggestions, and (iii) a compact, task-focused model can meet the cost and latency constraints of interactive completion. In the paper, we describe an end-to-end industrial pipeline for producing contextualized in-editor completion: disciplined data governance, multi-stage training that includes fill-in-the-middle and project context via supervised fine-tuning, and alignment via direct preference optimization using feedback from real-world scenarios. Our quality evaluations include both large-scale offline benchmarks and online telemetry from production deployments in JetBrains IDEs. Mellums are released under the Apache-2.0 license on HuggingFace, with a public model card providing a reproducible reference for practitioners. Our experience offers a pragmatic blueprint for taking a focused, open model from a research prototype to at scale production for hundreds of thousands of users.

Nikita Pavlichenko, Dmitry Ustalov

Today, comprehensive evaluation of large-scale machine learning models is possible thanks to the open datasets produced using crowdsourcing, such as SQuAD, MS COCO, ImageNet, SuperGLUE, etc. These datasets capture objective responses, assuming the single correct answer, which does not allow to capture the subjective human perception. In turn, pairwise comparison tasks, in which one has to choose between only two options, allow taking peoples' preferences into account for very challenging artificial intelligence tasks, such as information retrieval and recommender system evaluation. Unfortunately, the available datasets are either small or proprietary, slowing down progress in gathering better feedback from human users. In this paper, we present IMDB-WIKI-SbS, a new large-scale dataset for evaluating pairwise comparisons. It contains 9,150 images appearing in 250,249 pairs annotated on a crowdsourcing platform. Our dataset has balanced distributions of age and gender using the well-known IMDB-WIKI dataset as ground truth. We describe how our dataset is built and then compare several baseline methods, indicating its suitability for model evaluation.

Nikita Pavlichenko, Ivan Stelmakh, Dmitry Ustalov

Domain-specific data is the crux of the successful transfer of machine learning systems from benchmarks to real life. In simple problems such as image classification, crowdsourcing has become one of the standard tools for cheap and time-efficient data collection: thanks in large part to advances in research on aggregation methods. However, the applicability of crowdsourcing to more complex tasks (e.g., speech recognition) remains limited due to the lack of principled aggregation methods for these modalities. The main obstacle towards designing aggregation methods for more advanced applications is the absence of training data, and in this work, we focus on bridging this gap in speech recognition. For this, we collect and release CrowdSpeech -- the first publicly available large-scale dataset of crowdsourced audio transcriptions. Evaluation of existing and novel aggregation methods on our data shows room for improvement, suggesting that our work may entail the design of better algorithms. At a higher level, we also contribute to the more general challenge of developing the methodology for reliable data collection via crowdsourcing. In that, we design a principled pipeline for constructing datasets of crowdsourced audio transcriptions in any novel domain. We show its applicability on an under-resourced language by constructing VoxDIY -- a counterpart of CrowdSpeech for the Russian language. We also release the code that allows a full replication of our data collection pipeline and share various insights on best practices of data collection via crowdsourcing.

Varvara Logacheva, Denis Teslenko, Artem Shelmanov et al.

Disambiguation of word senses in context is easy for humans, but is a major challenge for automatic approaches. Sophisticated supervised and knowledge-based models were developed to solve this task. However, (i) the inherent Zipfian distribution of supervised training instances for a given word and/or (ii) the quality of linguistic knowledge representations motivate the development of completely unsupervised and knowledge-free approaches to word sense disambiguation (WSD). They are particularly useful for under-resourced languages which do not have any resources for building either supervised and/or knowledge-based models. In this paper, we present a method that takes as input a standard pre-trained word embedding model and induces a fully-fledged word sense inventory, which can be used for disambiguation in context. We use this method to induce a collection of sense inventories for 158 languages on the basis of the original pre-trained fastText word embeddings by Grave et al. (2018), enabling WSD in these languages. Models and system are available online.

Saba Anwar, Dmitry Ustalov, Nikolay Arefyev et al.

We present our system for semantic frame induction that showed the best performance in Subtask B.1 and finished as the runner-up in Subtask A of the SemEval 2019 Task 2 on unsupervised semantic frame induction (QasemiZadeh et al., 2019). Our approach separates this task into two independent steps: verb clustering using word and their context embeddings and role labeling by combining these embeddings with syntactical features. A simple combination of these steps shows very competitive results and can be extended to process other datasets and languages.

Dmitry Ustalov, Alexander Panchenko, Chris Biemann et al.

In this paper, we show how unsupervised sense representations can be used to improve hypernymy extraction. We present a method for extracting disambiguated hypernymy relationships that propagates hypernyms to sets of synonyms (synsets), constructs embeddings for these sets, and establishes sense-aware relationships between matching synsets. Evaluation on two gold standard datasets for English and Russian shows that the method successfully recognizes hypernymy relationships that cannot be found with standard Hearst patterns and Wiktionary datasets for the respective languages.

Dmitry Ustalov, Alexander Panchenko, Chris Biemann et al.

We present a detailed theoretical and computational analysis of the Watset meta-algorithm for fuzzy graph clustering, which has been found to be widely applicable in a variety of domains. This algorithm creates an intermediate representation of the input graph that reflects the "ambiguity" of its nodes. Then, it uses hard clustering to discover clusters in this "disambiguated" intermediate graph. After outlining the approach and analyzing its computational complexity, we demonstrate that Watset shows competitive results in three applications: unsupervised synset induction from a synonymy graph, unsupervised semantic frame induction from dependency triples, and unsupervised semantic class induction from a distributional thesaurus. Our algorithm is generic and can be also applied to other networks of linguistic data.

Dmitry Ustalov, Alexander Panchenko, Andrei Kutuzov et al.

We use dependency triples automatically extracted from a Web-scale corpus to perform unsupervised semantic frame induction. We cast the frame induction problem as a triclustering problem that is a generalization of clustering for triadic data. Our replicable benchmarks demonstrate that the proposed graph-based approach, Triframes, shows state-of-the art results on this task on a FrameNet-derived dataset and performing on par with competitive methods on a verb class clustering task.

Dmitry Ustalov, Denis Teslenko, Alexander Panchenko et al.

In this paper, we present Watasense, an unsupervised system for word sense disambiguation. Given a sentence, the system chooses the most relevant sense of each input word with respect to the semantic similarity between the given sentence and the synset constituting the sense of the target word. Watasense has two modes of operation. The sparse mode uses the traditional vector space model to estimate the most similar word sense corresponding to its context. The dense mode, instead, uses synset embeddings to cope with the sparsity problem. We describe the architecture of the present system and also conduct its evaluation on three different lexical semantic resources for Russian. We found that the dense mode substantially outperforms the sparse one on all datasets according to the adjusted Rand index.

Alexander Panchenko, Natalia Loukachevitch, Dmitry Ustalov et al.

The paper gives an overview of the Russian Semantic Similarity Evaluation (RUSSE) shared task held in conjunction with the Dialogue 2015 conference. There exist a lot of comparative studies on semantic similarity, yet no analysis of such measures was ever performed for the Russian language. Exploring this problem for the Russian language is even more interesting, because this language has features, such as rich morphology and free word order, which make it significantly different from English, German, and other well-studied languages. We attempt to bridge this gap by proposing a shared task on the semantic similarity of Russian nouns. Our key contribution is an evaluation methodology based on four novel benchmark datasets for the Russian language. Our analysis of the 105 submissions from 19 teams reveals that successful approaches for English, such as distributional and skip-gram models, are directly applicable to Russian as well. On the one hand, the best results in the contest were obtained by sophisticated supervised models that combine evidence from different sources. On the other hand, completely unsupervised approaches, such as a skip-gram model estimated on a large-scale corpus, were able score among the top 5 systems.

Alexander Panchenko, Anastasiya Lopukhina, Dmitry Ustalov et al.

The paper describes the results of the first shared task on word sense induction (WSI) for the Russian language. While similar shared tasks were conducted in the past for some Romance and Germanic languages, we explore the performance of sense induction and disambiguation methods for a Slavic language that shares many features with other Slavic languages, such as rich morphology and virtually free word order. The participants were asked to group contexts of a given word in accordance with its senses that were not provided beforehand. For instance, given a word "bank" and a set of contexts for this word, e.g. "bank is a financial institution that accepts deposits" and "river bank is a slope beside a body of water", a participant was asked to cluster such contexts in the unknown in advance number of clusters corresponding to, in this case, the "company" and the "area" senses of the word "bank". For the purpose of this evaluation campaign, we developed three new evaluation datasets based on sense inventories that have different sense granularity. The contexts in these datasets were sampled from texts of Wikipedia, the academic corpus of Russian, and an explanatory dictionary of Russian. Overall, 18 teams participated in the competition submitting 383 models. Multiple teams managed to substantially outperform competitive state-of-the-art baselines from the previous years based on sense embeddings.

Alexander Panchenko, Dmitry Ustalov, Stefano Faralli et al.

In this paper, we show how distributionally-induced semantic classes can be helpful for extracting hypernyms. We present methods for inducing sense-aware semantic classes using distributional semantics and using these induced semantic classes for filtering noisy hypernymy relations. Denoising of hypernyms is performed by labeling each semantic class with its hypernyms. On the one hand, this allows us to filter out wrong extractions using the global structure of distributionally similar senses. On the other hand, we infer missing hypernyms via label propagation to cluster terms. We conduct a large-scale crowdsourcing study showing that processing of automatically extracted hypernyms using our approach improves the quality of the hypernymy extraction in terms of both precision and recall. Furthermore, we show the utility of our method in the domain taxonomy induction task, achieving the state-of-the-art results on a SemEval'16 task on taxonomy induction.

Alexander Panchenko, Dmitry Ustalov, Nikolay Arefyev et al.

Semantic relatedness of terms represents similarity of meaning by a numerical score. On the one hand, humans easily make judgments about semantic relatedness. On the other hand, this kind of information is useful in language processing systems. While semantic relatedness has been extensively studied for English using numerous language resources, such as associative norms, human judgments, and datasets generated from lexical databases, no evaluation resources of this kind have been available for Russian to date. Our contribution addresses this problem. We present five language resources of different scale and purpose for Russian semantic relatedness, each being a list of triples (word_i, word_j, relatedness_ij). Four of them are designed for evaluation of systems for computing semantic relatedness, complementing each other in terms of the semantic relation type they represent. These benchmarks were used to organize a shared task on Russian semantic relatedness, which attracted 19 teams. We use one of the best approaches identified in this competition to generate the fifth high-coverage resource, the first open distributional thesaurus of Russian. Multiple evaluations of this thesaurus, including a large-scale crowdsourcing study involving native speakers, indicate its high accuracy.

Dmitry Ustalov, Mikhail Chernoskutov, Chris Biemann et al.

Graph-based synset induction methods, such as MaxMax and Watset, induce synsets by performing a global clustering of a synonymy graph. However, such methods are sensitive to the structure of the input synonymy graph: sparseness of the input dictionary can substantially reduce the quality of the extracted synsets. In this paper, we propose two different approaches designed to alleviate the incompleteness of the input dictionaries. The first one performs a pre-processing of the graph by adding missing edges, while the second one performs a post-processing by merging similar synset clusters. We evaluate these approaches on two datasets for the Russian language and discuss their impact on the performance of synset induction methods. Finally, we perform an extensive error analysis of each approach and discuss prominent alternative methods for coping with the problem of the sparsity of the synonymy dictionaries.

Alexander Panchenko, Fide Marten, Eugen Ruppert et al.

Interpretability of a predictive model is a powerful feature that gains the trust of users in the correctness of the predictions. In word sense disambiguation (WSD), knowledge-based systems tend to be much more interpretable than knowledge-free counterparts as they rely on the wealth of manually-encoded elements representing word senses, such as hypernyms, usage examples, and images. We present a WSD system that bridges the gap between these two so far disconnected groups of methods. Namely, our system, providing access to several state-of-the-art WSD models, aims to be interpretable as a knowledge-based system while it remains completely unsupervised and knowledge-free. The presented tool features a Web interface for all-word disambiguation of texts that makes the sense predictions human readable by providing interpretable word sense inventories, sense representations, and disambiguation results. We provide a public API, enabling seamless integration.

Dmitry Ustalov, Nikolay Arefyev, Chris Biemann et al.

We present a new approach to extraction of hypernyms based on projection learning and word embeddings. In contrast to classification-based approaches, projection-based methods require no candidate hyponym-hypernym pairs. While it is natural to use both positive and negative training examples in supervised relation extraction, the impact of negative examples on hypernym prediction was not studied so far. In this paper, we show that explicit negative examples used for regularization of the model significantly improve performance compared to the state-of-the-art approach of Fu et al. (2014) on three datasets from different languages.

Dmitry Ustalov, Alexander Panchenko, Chris Biemann

This paper presents a new graph-based approach that induces synsets using synonymy dictionaries and word embeddings. First, we build a weighted graph of synonyms extracted from commonly available resources, such as Wiktionary. Second, we apply word sense induction to deal with ambiguous words. Finally, we cluster the disambiguated version of the ambiguous input graph into synsets. Our meta-clustering approach lets us use an efficient hard clustering algorithm to perform a fuzzy clustering of the graph. Despite its simplicity, our approach shows excellent results, outperforming five competitive state-of-the-art methods in terms of F-score on three gold standard datasets for English and Russian derived from large-scale manually constructed lexical resources.

Dmitry Ustalov

Linguistic resources can be populated with data through the use of such approaches as crowdsourcing and gamification when motivated people are involved. However, current crowdsourcing genre taxonomies lack the concept of cooperation, which is the principal element of modern video games and may potentially drive the annotators' interest. This survey on crowdsourcing taxonomies and cooperation in linguistic resources provides recommendations on using cooperation in existent genres of crowdsourcing and an evidence of the efficiency of cooperation using a popular Russian linguistic resource created through crowdsourcing as an example.