Search Results for author: Hady Elsahar

Found 15 papers, 7 papers with code

Unsupervised and Distributional Detection of Machine-Generated Text

no code implementations4 Nov 2021 Matthias Gallé, Jos Rozen, Germán Kruszewski, Hady Elsahar

We propose a method to detect those machine-generated documents leveraging repeated higher-order n-grams, which we show over-appear in machine-generated text as compared to human ones.

Text Generation

Energy-Based Models for Code Generation under Compilability Constraints

no code implementations9 Jun 2021 Tomasz Korbak, Hady Elsahar, Marc Dymetman, Germán Kruszewski

Neural language models can be successfully trained on source code, leading to applications such as code completion.

Code Completion Code Generation

Symbol-Shift Equivariant Neural Networks

no code implementations1 Jan 2021 David Salinas, Hady Elsahar

Neural networks have been shown to have poor compositionality abilities: while they can produce sophisticated output given sufficient data, they perform patchy generalization and fail to generalize to new symbols (e. g. switching a name in a sentence by a less frequent one or one not seen yet).

Question Answering Translation

A Distributional Approach to Controlled Text Generation

1 code implementation ICLR 2021 Muhammad Khalifa, Hady Elsahar, Marc Dymetman

From that optimal representation we then train a target controlled Autoregressive LM through an adaptive distributional variant of Policy Gradient.

Text Generation

Self-Supervised and Controlled Multi-Document Opinion Summarization

no code implementations EACL 2021 Hady Elsahar, Maximin Coavoux, Matthias Gallé, Jos Rozen

We address the problem of unsupervised abstractive summarization of collections of user generated reviews with self-supervision and control.

Abstractive Text Summarization

To Annotate or Not? Predicting Performance Drop under Domain Shift

no code implementations IJCNLP 2019 Hady Elsahar, Matthias Gall{\'e}

In this paper, we study the problem of predicting the performance drop of modern NLP models under domain-shift, in the absence of any target domain labels.

General Classification POS +1

Zero-Shot Question Generation from Knowledge Graphs for Unseen Predicates and Entity Types

1 code implementation NAACL 2018 Hady Elsahar, Christophe Gravier, Frederique Laforest

We present a neural model for question generation from knowledge base triples in a "Zero-Shot" setup, that is generating questions for triples containing predicates, subject types or object types that were not seen at training time.

Knowledge Graphs Question Generation

Unsupervised Open Relation Extraction

1 code implementation22 Jan 2018 Hady Elsahar, Elena Demidova, Simon Gottschalk, Christophe Gravier, Frederique Laforest

We explore methods to extract relations between named entities from free text in an unsupervised setting.

Relation Extraction Word Embeddings

High Recall Open IE for Relation Discovery

no code implementations IJCNLP 2017 Hady Elsahar, Christophe Gravier, Frederique Laforest

Relation Discovery discovers predicates (relation types) from a text corpus relying on the co-occurrence of two named entities in the same sentence.

Open Information Extraction Relation Extraction

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