Search results for “BERT”

Real-time monitoring of scientific papers and technological news requires fast processing of complicated search demands motivated by thematically relevant information acquisition. For this case, the authors develop an exploratory search engine based on probabilistic hierarchical topic modeling. Topic model gives a low dimensional sparse interpretable vector representation (topical embedding) of a text, which is used for ranking documents by their similarity to the query. They explore several ways of comparing topical vectors including searching with thematically homogeneous text segments. Topical hierarchies are built using the regularized EM-algorithm from BigARTM project. The topic-based search achieves better precision and recall than other approaches (TF-IDF, fastText, LSTM, BERT) and even human assessors who spend up to an hour to complete the same search task. They also discover that blending hierarchical topic vectors with neural pretrained embeddings is a promising way of enriching both models that helps to get precision and recall higher than 90

Machine translation is a natural language text processing task that aims to automatically translate input text from one language into another language. The currently known machine translation models show a fairly high quality of translation between large languages, but for smaller language areas, represented by less data, the problem is still not solved. Different methods are used to deal with various errors in automatic translation systems. This paper discusses approaches that use translation models of reverse language directions and improve consistency between translations of the same text using direct and reverse translation models. The paper presents a general theoretical justification for such methods in terms of solving the likelihood maximization problem and also proposes a method for stable training of modern models using cyclic translations.

Three general methods for obtaining exact bounds on the probability of overfitting are proposed within statistical learning theory: a method of generating and destroying sets, a recurrent method, and a blockwise method. Six particular cases are considered to illustrate the application of these methods. These are the following model sets of predictors: a pair of predictors, a layer of a Boolean cube, an interval of a Boolean cube, a monotonic chain, a unimodal chain, and a unit neighborhood of the best predictor. For the interval and the unimodal chain, the results of numerical experiments are presented that demonstrate the effects of splitting and similarity on the probability of overfitting.

Social studies of the Internet have adopted large-scale text mining for unsupervised discovery of topics related to specific subjects. A recently developed approach to topic modeling, additive regularization of topic models (ARTM), provides fast inference and more control over the topics with a wide variety of possible regularizers than developing LDA extensions. We apply ARTM to mining ethnic-related content from Russian-language blogosphere, introduce a new combined regularizer, and compare models derived from ARTM with LDA. We show with human evaluations that ARTM is better for mining topics on specific subjects, finding more relevant topics of higher or comparable quality.

We propose a generalized probabilistic topic model of text corpora which can incorporate heuristics of Bayesian regularization, sampling, frequent parameters update, and robustness in any combinations. Wellknown models PLSA, LDA, CVB0, SWB, and many others can be considered as special cases of the proposed broad family of models. We propose the robust PLSA model and show that it is more sparse and performs better that regularized models like LDA.

Sharing scientific knowledge in the community is an important endeavor. However, most papers are written in English, which makes dissemination of knowledge in countries where English is not spoken by the majority of people harder. Nowadays, machine translation and language models may help to solve this problem, but it is still complicated to train and evaluate models in languages other than English with no or little data in the required language. To address this, we propose the first benchmark for evaluating models on scientific texts in Russian. It consists of papers from Russian electronic library of scientific publications. We also present a set of tasks which can be used to fine-tune various models on our data and provide a detailed comparison between state-of-the-art models on our benchmark.