Search Results for author: Aliaksei Severyn

Found 33 papers, 12 papers with code

Teaching Small Language Models to Reason

no code implementations16 Dec 2022 Lucie Charlotte Magister, Jonathan Mallinson, Jakub Adamek, Eric Malmi, Aliaksei Severyn

Chain of thought prompting successfully improves the reasoning capabilities of large language models, achieving state of the art results on a range of datasets.

GSM8K Knowledge Distillation

Text Generation with Text-Editing Models

no code implementations NAACL (ACL) 2022 Eric Malmi, Yue Dong, Jonathan Mallinson, Aleksandr Chuklin, Jakub Adamek, Daniil Mirylenka, Felix Stahlberg, Sebastian Krause, Shankar Kumar, Aliaksei Severyn

Text-editing models have recently become a prominent alternative to seq2seq models for monolingual text-generation tasks such as grammatical error correction, simplification, and style transfer.

Grammatical Error Correction Style Transfer +1

EdiT5: Semi-Autoregressive Text-Editing with T5 Warm-Start

no code implementations24 May 2022 Jonathan Mallinson, Jakub Adamek, Eric Malmi, Aliaksei Severyn

This is achieved by decomposing the generation process into three sub-tasks: (1) tagging to decide on the subset of input tokens to be preserved in the output, (2) re-ordering to define their order in the output text, and (3) insertion to infill the missing tokens that are not present in the input.

Grammatical Error Correction Sentence Fusion

Controlled Text Generation as Continuous Optimization with Multiple Constraints

1 code implementation NeurIPS 2021 Sachin Kumar, Eric Malmi, Aliaksei Severyn, Yulia Tsvetkov

As large-scale language model pretraining pushes the state-of-the-art in text generation, recent work has turned to controlling attributes of the text such models generate.

Language Modelling Machine Translation +3

A Simple Recipe for Multilingual Grammatical Error Correction

1 code implementation ACL 2021 Sascha Rothe, Jonathan Mallinson, Eric Malmi, Sebastian Krause, Aliaksei Severyn

This paper presents a simple recipe to train state-of-the-art multilingual Grammatical Error Correction (GEC) models.

 Ranked #1 on Grammatical Error Correction on Falko-MERLIN (using extra training data)

Grammatical Error Correction

Unsupervised Text Style Transfer with Padded Masked Language Models

no code implementations EMNLP 2020 Eric Malmi, Aliaksei Severyn, Sascha Rothe

This allows us to identify the source tokens to delete to transform the source text to match the style of the target domain.

Sentence Fusion Style Transfer +2

Felix: Flexible Text Editing Through Tagging and Insertion

2 code implementations Findings of the Association for Computational Linguistics 2020 Jonathan Mallinson, Aliaksei Severyn, Eric Malmi, Guillermo Garrido

We achieve this by decomposing the text-editing task into two sub-tasks: tagging to decide on the subset of input tokens and their order in the output text and insertion to in-fill the missing tokens in the output not present in the input.

Automatic Post-Editing Language Modelling +3

Breaking the Softmax Bottleneck via Learnable Monotonic Pointwise Non-linearities

no code implementations21 Feb 2019 Octavian-Eugen Ganea, Sylvain Gelly, Gary Bécigneul, Aliaksei Severyn

The Softmax function on top of a final linear layer is the de facto method to output probability distributions in neural networks.

Language Modelling Text Generation

Prosody Modifications for Question-Answering in Voice-Only Settings

2 code implementations11 Jun 2018 Aleksandr Chuklin, Aliaksei Severyn, Johanne Trippas, Enrique Alfonseca, Hanna Silen, Damiano Spina

Many popular form factors of digital assistants---such as Amazon Echo, Apple Homepod, or Google Home---enable the user to hold a conversation with these systems based only on the speech modality.

Informativeness Question Answering

Eval all, trust a few, do wrong to none: Comparing sentence generation models

no code implementations21 Apr 2018 Ondřej Cífka, Aliaksei Severyn, Enrique Alfonseca, Katja Filippova

In this paper, we study recent neural generative models for text generation related to variational autoencoders.

Text Generation

Learning to Learn from Weak Supervision by Full Supervision

1 code implementation30 Nov 2017 Mostafa Dehghani, Aliaksei Severyn, Sascha Rothe, Jaap Kamps

In this paper, we propose a method for training neural networks when we have a large set of data with weak labels and a small amount of data with true labels.

Neural Ranking Models with Weak Supervision

1 code implementation28 Apr 2017 Mostafa Dehghani, Hamed Zamani, Aliaksei Severyn, Jaap Kamps, W. Bruce Croft

Our experiments indicate that employing proper objective functions and letting the networks to learn the input representation based on weakly supervised data leads to impressive performance, with over 13% and 35% MAP improvements over the BM25 model on the Robust and the ClueWeb collections.

Ad-Hoc Information Retrieval Information Retrieval +1

A Hybrid Convolutional Variational Autoencoder for Text Generation

3 code implementations EMNLP 2017 Stanislau Semeniuta, Aliaksei Severyn, Erhardt Barth

In this paper we explore the effect of architectural choices on learning a Variational Autoencoder (VAE) for text generation.

Language Modelling Text Generation

Modeling Relational Information in Question-Answer Pairs with Convolutional Neural Networks

no code implementations5 Apr 2016 Aliaksei Severyn, Alessandro Moschitti

In this paper, we propose convolutional neural networks for learning an optimal representation of question and answer sentences.

Globally Normalized Transition-Based Neural Networks

1 code implementation ACL 2016 Daniel Andor, Chris Alberti, David Weiss, Aliaksei Severyn, Alessandro Presta, Kuzman Ganchev, Slav Petrov, Michael Collins

Our model is a simple feed-forward neural network that operates on a task-specific transition system, yet achieves comparable or better accuracies than recurrent models.

Dependency Parsing Part-Of-Speech Tagging +1

Recurrent Dropout without Memory Loss

2 code implementations COLING 2016 Stanislau Semeniuta, Aliaksei Severyn, Erhardt Barth

This paper presents a novel approach to recurrent neural network (RNN) regularization.

SenTube: A Corpus for Sentiment Analysis on YouTube Social Media

no code implementations LREC 2014 Olga Uryupina, Barbara Plank, Aliaksei Severyn, Agata Rotondi, Aless Moschitti, ro

In this paper we present SenTube -- a dataset of user-generated comments on YouTube videos annotated for information content and sentiment polarity.

Document Classification Informativeness +3

Cannot find the paper you are looking for? You can Submit a new open access paper.