We introduce KazQAD—a Kazakh open-domain question answering (ODQA) dataset—that can be used in both reading comprehension and full ODQA settings, as well as for information retrieval experiments. KazQAD contains just under 6,000 unique questions with extracted short answers and nearly 12,000 passage-level relevance judgements. We use a combination of machine translation, Wikipedia search, and in-house manual annotation to ensure annotation efficiency and data quality. The questions come from two sources: translated items from the Natural Questions (NQ) dataset (only for training) and the original Kazakh Unified National Testing (UNT) exam (for development and testing). The accompanying text corpus contains more than 800,000 passages from the Kazakh Wikipedia. As a supplementary dataset, we release around 61,000 question-passage-answer triples from the NQ dataset that have been machine-translated into Kazakh. We develop baseline retrievers and readers that achieve reasonable scores in retrieval (NDCG10 = 0.389 MRR = 0.382), reading comprehension (EM = 38.5 F1 = 54.2), and full ODQA (EM = 17.8 F1 = 28.7) settings. Nevertheless, these results are substantially lower than state-of-the-art results for English QA collections, and we think that there should still be ample room for improvement. We also show that the current OpenAI’s ChatGPTv3.5 is not able to answer KazQAD test questions in the closed-book setting with acceptable quality. The dataset is freely available under the Creative Commons licence (CC BY-SA) at url https://github.com/IS2AI/KazQAD

In this study, we focus on automatic humor detection, a highly relevant task for conversational AI.To date, there are several English datasets for this task, but little research on how models trained on them generalize and behave in the wild.To fill this gap, we carefully analyze existing datasets, train RoBERTabased and Naïve Bayes classifiers on each of them, and test on the rest.Training and testing on the same dataset yields good results, but the transferability of the models varies widely.Models trained on datasets with jokes from different sources show better transferability, while the amount of training data has a smaller impact.The behavior of the models on out-of-domain data is unstable, suggesting that some of the models overfit, while others learn non-specific humor characteristics.An adversarial attack shows that models trained on pun datasets are less robust.We also evaluate the sense of humor of the chatGPT and Flan-UL2 models in a zero-shot scenario.The LLMs demonstrate competitive results on humor datasets and a more stable behavior on out-of-domain data.We believe that the obtained results will facilitate the development of new datasets and evaluation methodologies in the field of computational humor.We've made all the data from the study and the trained models publicly available: https://github.com/ Humor-Research/Humor-detection.Tell me which are funny, which are not -and which get a giggle first time but are cold pancakes without honey to hear twice.

MOTIVATION: This article describes NEREL-BIO-an annotation scheme and corpus of PubMed abstracts in Russian and smaller number of abstracts in English. NEREL-BIO extends the general domain dataset NEREL by introducing domain-specific entity types. NEREL-BIO annotation scheme covers both general and biomedical domains making it suitable for domain transfer experiments. NEREL-BIO provides annotation for nested named entities as an extension of the scheme employed for NEREL. Nested named entities may cross entity boundaries to connect to shorter entities nested within longer entities, making them harder to detect. RESULTS: NEREL-BIO contains annotations for 700+ Russian and 100+ English abstracts. All English PubMed annotations have corresponding Russian counterparts. Thus, NEREL-BIO comprises the following specific features: annotation of nested named entities, it can be used as a benchmark for cross-domain (NEREL → NEREL-BIO) and cross-language (English → Russian) transfer. We experiment with both transformer-based sequence models and machine reading comprehension models and report their results. AVAILABILITY AND IMPLEMENTATION: The dataset and annotation guidelines are freely available at https://github.com/nerel-ds/NEREL-BIO.

At least 5% of questions submitted to search engines ask about cause-effect relationships in some way. To support the development of tailored approaches that can answer such questions, we construct Webis-CausalQA-22, a benchmark corpus of 1.1 million causal questions with answers. We distinguish different types of causal questions using a novel typology derived from a data-driven, manual analysis of questions from ten large question answering (QA) datasets. Using high-precision lexical rules, we extract causal questions of each type from these datasets to create our corpus. As an initial baseline, the state-of-the-art QA model UnifiedQA achieves a ROUGE-L F1 score of 0.48 on our new benchmark.

In this paper, we describe entity linking annotation over nested named entities in the recently released Russian NEREL dataset for information extraction. The NEREL collection is currently the largest Russian dataset annotated with entities and relations. It includes 933 news texts with annotation of 29 entity types and 49 relation types. The paper describes the main design principles behind NEREL’s entity linking annotation, provides its statistics, and reports evaluation results for several entity linking baselines. To date, 38,152 entity mentions in 933 documents are linked to Wikidata. The NEREL dataset is publicly available.

We present an approach to identify argumentative questions among web search queries. Argumentative questions ask for reasons to support a certain stance on a controversial topic, such as ''Should marijuana be legalized?'' Controversial topics entail opposing stances, and hence can be supported or opposed by various arguments. Argumentative questions pose a challenge for search engines since they should be answered with both pro and con arguments in order to not bias a user toward a certain stance.

In this paper, we analyze comparative questions and answers. At least 3%~of the questions submitted to search engines are comparative; ranging from simple facts like "Did Messi or Ronaldo score more goals in 2021?'' to life-changing and probably highly subjective questions like "Is it better to move abroad or stay?''. Ideally, answers to subjective comparative questions would reflect diverse opinions so that the asker can come to a well-informed decision. To better understand the information needs behind comparative questions, we develop approaches to extract the mentioned comparison objects and aspects. As a first step to answer comparative questions, we develop an approach that detects the stances of potential result nuggets (i.e., text passages containing the comparison objects). Our approaches are trained and evaluated on a set of 31,000~English questions from existing datasets that we label as comparative or not. In the 3,500~comparative questions, we label the comparison objects, aspects, and predicates. For 950~questions, we collect answers from online forums and label the stance towards the comparison objects. In the experiments, our approaches recall~71% of the comparative questions with a perfect precision of~1.0, recall~92% of subjective comparative questions with a precision of~0.98, and identify the comparison objects and aspects with an F1 of~0.93 and~0.80, respectively. The stance detector fine-tuned on pairs of objects and answers achieves an accuracy of~0.63.

Quality of search engine results returned to health-related questions is very critical, since a searcher may directly trust any suggestion in the top results. We analyze search questions that mention diseases / symptoms and remedies that are potential health-related misbeliefs. Using lists of medical and alternative medicine terms, we extract health-related search questions from 1.5~billion questions submitted to Yandex. As an initial study, we sample 30 frequent questions that contain a disease--remedy pair like "Can hepatitis be cured with milk thistle?". For each question, we carefully identify a ground truth answer in the medical literature and annotate the top-10 Yandex search result snippets as confirming the belief, rejecting it, or giving no answer. Our analysis shows that about 44%~of the snippets (that users may simply interpret as definitive answers!) confirm some untrue beliefs and are wrong, and only few include health risk warnings about using toxic plants.

The paper describes the second version of RuBQ, a Russian dataset for knowledge base question answering (KBQA) over Wikidata. Whereas the first version builds on Q&A pairs harvested online, the extension is based on questions obtained through search engine query suggestion services. The questions underwent crowdsourced and in-house annotation in a quite different fashion compared to the first edition. The dataset doubled in size: RuBQ 2.0 contains 2,910 questions along with the answers and SPARQL queries. The dataset also incorporates answer-bearing paragraphs from Wikipedia for the majority of questions. The dataset is suitable for the evaluation of KBQA, machine reading comprehension (MRC), hybrid questions answering, as well as semantic parsing. We provide the analysis of the dataset and report several KBQA and MRC baseline results. The dataset is freely available under the CC-BY-4.0 license.

Due to high annotation costs making the best use of existing human-created training data is an important research direction. We, therefore, carry out a systematic evaluation of transferability of BERT-based neural ranking models across five English datasets. Previous studies focused primarily on zero-shot and few-shot transfer from a large dataset to a dataset with a small number of queries. In contrast, each of our collections has a substantial number of queries, which enables a full-shot evaluation mode and improves reliability of our results. Furthermore, since source datasets licences often prohibit commercial use, we compare transfer learning to training on pseudo-labels generated by a BM25 scorer. We find that training on pseudo-labels---possibly with subsequent fine-tuning using a modest number of annotated queries---can produce a competitive or better model compared to transfer learning. Yet, it is necessary to improve the stability and/or effectiveness of the few-shot training, which, sometimes, can degrade performance of a pretrained model.

\beginabstract We analyze comparative questions, i.e., questions asking to compare different items, that were submitted to Yandex in 2012. Responses to such questions might be quite different from the simple "ten blue links'' and could, for example, aggregate pros and cons of the different options as direct answers. However, changing the result presentation is an intricate decision such that the classification of comparative questions forms a highly precision-oriented task.

The task of humor recognition has attracted a lot of attention recently due to the urge to process large amounts of user-generated texts and rise of conversational agents.We collected a dataset of jokes and funny dialogues in Russian from various online resources and complemented them carefully with unfunny texts with similar lexical properties.The dataset comprises of more than 300,000 short texts, which is significantly larger than any previous humor-related corpus.Manual annotation of about 2,000 items proved the reliability of the corpus construction approach.Further, we applied language model fine-tuning for text classification and obtained an F1 score of 0.91 on test set, which constitutes a considerable gain over baseline methods.The dataset is freely available for research community.

Nowadays natural language user interfaces, such as chatbots and conversational agents, are very common. A desirable trait of such applications is a sense of humor. It is, therefore, important to be able to measure quality of humorous responses. However, humor evaluation is hard since humor is highly subjective. To address this problem, we conducted an online evaluation of 30 dialog jokes from different sources by almost 300 participants -- volunteers and Mechanical Turk workers. We collected joke ratings along with participants» age, gender, and language proficiency. Results show that demographics and joke topics can partly explain variation in humor judgments. We expect that these insights will aid humor evaluation and interpretation. The findings can also be of interest for humor generation methods in conversational systems.

User system trust is critical to the uptake of recommendations, and several factors of trust have been identified and compared. In this paper we present a cross-cultural, crowdsourced study examining user perceptions of nine factors of trust and link the observed differences to trust development processes and cultural dimensions. While some factors consistently instil trust, others are preferred only in certain countries. Our findings and the discovered links are important for design of trusted recommender systems.

Search as a dialogue is an emerging paradigm that is fueled by the proliferation of mobile devices and technological advances, e.g. in speech recognition and natural language processing. Such an interface allows search systems to engage in a dialogue with users aimed at fulfilling their information needs. One key capability required to make such search dialogues effective is asking clarification questions (CLARQ) proactively, when a user's intent is not clear, which could help the system provide more useful responses. With this in mind, we explore the dialogues between the users on a community question answering (CQA) website as a rich repository of information-seeking interactions. In particular, we study the clarification questions asked by CQA users in two different domains, analyze their behavior, and the types of clarification questions asked. Our results suggest that the types of CLARQ are very diverse, while the questions themselves tend to be specific and require both domain- and general knowledge. However, focusing on popular CLARQ types and domains can be fruitful. As a first step towards automatic generation of clarification questions, we explore the problem of predicting the specific subject of a clarification question. Our findings can be useful for future improvements of intelligent dialog search and question answering systems.

YARN (Yet Another RussNet), a project started in 2013, aims at creating a large open WordNet-like thesaurus for Russian by means of crowdsourcing.The first stage of the project was to create noun synsets.Currently, the resource comprises 48K+ word entries and 44K+ synsets.More than 200 people have taken part in assembling synsets throughout the project.The paper describes the linguistic, technical, and organizational principles of the project, as well as the evaluation results, lessons learned, and the future plans.

We analyze the question queries submitted to a large commercial web search engine to get insights about what people ask, and to better tailor the search results to the users' needs. Based on a dataset of about one billion question queries submitted during the year 2012, we investigate askers' querying behavior with the support of automatic query categorization. While the importance of question queries is likely to increase, at present they only make up 3-4% of the total search traffic.

YARN (Yet Another RussNet) project started in 2013 aims at creating a large open thesaurus for Russian using crowdsourcing. This paper describes synset assembly interface developed within the project — motivation behind it, design, usage scenarios, implementation details, and first experimental results.

In this poster, we present a method for extracting queries related to real-life events, or news-related queries, from large web query logs. The method employs query frequencies and search over a collection of recent news. News-related queries can be helpful for disambiguating user information needs, as well as for effective online news processing. The performed evaluation proves that the method yields good precision.

The poster describes a fast, simple, yet accurate method to associate large amounts of web resources stored in a search engine database with geographic locations. The method uses location-by-IP data, domain names, and content-related features: ZIP and area codes. The novelty of the approach lies in building location-by-IP database by using continuous IP blocks method. Another contribution is domain name analysis. The method uses search engine infrastructure and makes it possible to effectively associate large amounts of search engine data with geography on a regular basis. Experiments ran on Yandex search engine index; evaluation has proved the efficacy of the approach.