Papers by André Freitas

32 papers
Building a Knowledge Graph from Natural Language Definitions for Interpretable Text Entailment Recognition (L18-1)

Copied to clipboard

Challenge: a conceptual model for dictionary definitions is used to construct a knowledge graph from natural language definitions.
Approach: They propose a method for automatically building a graph world knowledge base from natural language definitions.
Outcome: The proposed method was used in an interpretable text entailment recognition approach.
Estimating the Causal Effects of Natural Logic Features in Transformer-Based NLI Models (2024.lrec-main)

Copied to clipboard

Challenge: a number of studies have reported high accuracies in NLP tasks due to simple heuristics and dataset artifacts.
Approach: They use a case where two words/terms occur in a shared context to construct a causal diagram . they also investigate the robustness to irrelevant changes and sensitivity to impactful changes of Transformers .
Outcome: The proposed method bolsters the fact that similar benchmark accuracy scores may be observed for models that exhibit very different behaviour.
DBee: A Database for Creating and Managing Knowledge Graphs and Embeddings (D19-53)

Copied to clipboard

Challenge: DBee provides a data model which operates over knowledge graphs and embedding vector spaces .
Approach: They describe a database which provides a data model which exploits the semantic properties of large-scale knowledge graphs and embedding vector spaces.
Outcome: The proposed model exploits the semantic properties of both types of representations.
Graphene: a Context-Preserving Open Information Extraction System (C18-2)

Copied to clipboard

Challenge: Graphene is an open IE system that generates accurate, meaningful and complete propositions . current systems tend to extract propositions with long argument phrases that can be further decomposed into meaningful propositions, with each of them representing a separate fact.
Approach: They propose a lightweight Open IE system that generates accurate, meaningful propositions . they identify the rhetorical relations that hold between them to maintain their semantic relationship .
Outcome: The proposed system generates propositions that are accurate, meaningful and complete . it preserves the context of the relational tuples extracted from the source sentence .
Montague semantics and modifier consistency measurement in neural language models (2025.coling-main)

Copied to clipboard

Challenge: Existing studies on distributional language models have been focused on linguistics and their relationship with semantic formalisms for decades.
Approach: They propose a method for measuring compositional behavior in contemporary language embedding models by introducing three new tests inspired by Montague semantics.
Outcome: The proposed method measures compositional behavior in language embedding models on adjectival modifier phenomena in adjective-noun phrases.
Unification-based Reconstruction of Multi-hop Explanations for Science Questions (2021.eacl-main)

Copied to clipboard

Challenge: Existing approaches build explanations considering each question in isolation, but new approach leverages explanatory patterns emerging in scientific explanations.
Approach: They propose a framework for reconstructing multi-hop explanations in science Question Answering . they integrate lexical relevance with the notion of unification power to rank atomic facts .
Outcome: The proposed method achieves results competitive with Transformers, but is faster and scalable to large explanatory corpora.
Transforming Complex Sentences into a Semantic Hierarchy (P19-1)

Copied to clipboard

Challenge: Existing approaches for recursively splitting and rephrasing complex English sentences into a semantic hierarchy of simplified sentences are lacking.
Approach: They propose a method for recursively splitting and rephrasing complex English sentences into a semantic hierarchy of simplified sentences.
Outcome: The proposed approach outperforms state-of-the-art approaches in MT and information extraction tasks.
Identifying and Explaining Discriminative Attributes (D19-1)

Copied to clipboard

Challenge: Existing word vector representation models lack latent features (dense vectors) identifying discriminative attributes can motivate the development of word vector models with finer semantics.
Approach: They propose to use a word vector representation model to identify discriminative attributes by combining knowledge graphs with images to construct explicit vector spaces.
Outcome: The proposed model performs comparable to state-of-the-art systems while providing full model transparency and explainability.
Formal Semantic Controls over Language Models (2024.lrec-tutorials)

Copied to clipboard

Challenge: Text embeddings provide a concise representation of the semantics of sentences and larger spans of text, rather than individual words, capturing a wide range of linguistic features.
Approach: They propose to shorten the gap between latent semantics and formal symbolics by comparing distributional models to symbolic models grounded on formal linguistics and well-defined mathematical properties.
Outcome: This paper examines the analysis and control of text representations, covering methods from pooling to LLM-based.
PhysNLU: A Language Resource for Evaluating Natural Language Understanding and Explanation Coherence in Physics (2022.lrec-1)

Copied to clipboard

Challenge: physicists use mathematics to reason and explain, separates their field from other disciplines, including mathematics.
Approach: They present a dataset to evaluate the performance of language models in physics . they find that language models are challenged by coherence related tasks in physicists .
Outcome: The proposed models are able to perform well on coherence-related tasks even when trained on natural language objectives.
Does My Representation Capture X? Probe-Ably (2021.acl-demo)

Copied to clipboard

Challenge: Probing (or diagnostic classification) has become a popular strategy for investigating whether a given set of intermediate features is present in the representations of neural models.
Approach: They propose to use an extendable probing framework to automate the application of probing methods to the user’s inputs.
Outcome: The proposed framework automates the application of probing methods to the user’s inputs.
A Multilingual Test Collection for the Semantic Search of Entity Categories (L18-1)

Copied to clipboard

Challenge: Despite the high popularity of entity search, entity categories have not received equal attention.
Approach: They propose to make public a multilingual test collection comprehending English, Portuguese and German to meet the demands of the entity search community.
Outcome: The proposed test collection comprehends English, Portuguese and German and provides comparative baselines and an analysis of the results.
Introductory Tutorial: Reasoning with Natural Language Explanations (2024.emnlp-tutorials)

Copied to clipboard

Challenge: Existing paradigms for explanation-based NLIs lack a clear understanding of the nature of human reasoning.
Approach: They propose to use natural language explanations to build models that address downstream tasks through explicit construction of a natural language.
Outcome: In contrast to the existing paradigm based on Deep Learning, explanation-based NLI focuses on developing and evaluating models that address downstream tasks through the explicit construction of a natural language explanation.
Text Classification and Prediction in the Legal Domain (2022.lrec-1)

Copied to clipboard

Challenge: a case study combines text classification and legal judgment prediction for flight compensation . a human-in-the-loop model outperformed human prediction when predicting a claim being successful .
Approach: They combine transformer-based classification models with human-in-the-loop data to classify airlines' responses to flight compensation claims.
Outcome: The proposed models outperform human prediction when predicting a legal claim success.
Premise Selection in Natural Language Mathematical Texts (2020.acl-main)

Copied to clipboard

Challenge: Existing tasks for natural language premise selection are limited and difficult for humans to interpret and write.
Approach: They propose to use natural language premise selection task to predict premises that will be useful to prove a particular statement.
Outcome: The proposed approach improves the performance of baselines and multi-hop premise selection tasks.
Diff-Explainer: Differentiable Convex Optimization for Explainable Multi-hop Inference (2022.tacl-1)

Copied to clipboard

Challenge: Existing explainable multi-hop inference models are regarded as black-boxes due to their ability to transfer linguistic and semantic information to downstream tasks, posing concerns about interpretability and transparency of their predictions.
Approach: They propose a hybrid framework that integrates explicit constraints with neural architectures through differentiable convex optimization to answer and explain multi-hop questions in natural language.
Outcome: The proposed framework improves performance on scientific and commonsense QA tasks while still providing structured explanations in support of its predictions.
A Differentiable Integer Linear Programming Solver for Explanation-Based Natural Language Inference (2024.lrec-main)

Copied to clipboard

Challenge: Existing ILP frameworks are non-differentiable and cannot be integrated as part of a broader deep learning architecture.
Approach: They propose a neuro-symbolic architecture for explanation-based NLI based on DBCS.
Outcome: The proposed approach achieves superior performance when compared to existing solvers and black-box solver.
Introduction to Mathematical Language Processing: Informal Proofs, Word Problems, and Supporting Tasks (2023.tacl-1)

Copied to clipboard

Challenge: Using mathematical language processing methods, we analyze prevailing methods, existing limitations, and promising avenues for future research.
Approach: They analyze mathematical language processing methods from recent years and highlight prevailing methodologies, existing limitations and promising avenues for future research.
Outcome: The proposed methods highlight prevailing methods, existing limitations and promising avenues for future research.
Learning Disentangled Representations for Natural Language Definitions (2023.findings-eacl)

Copied to clipboard

Challenge: Disentangling the encodings of neural models is a fundamental aspect for improving interpretability, semantic control and downstream task performance in Natural Language Processing.
Approach: They propose to use syntactic and semantic regularities in textual data to provide models with both structural biases and generative factors.
Outcome: The proposed model outperforms baselines on several qualitative and quantitative benchmarks and improves the results in the downstream task of definition modeling.
Indra: A Word Embedding and Semantic Relatedness Server (L18-1)

Copied to clipboard

Challenge: Word embedding/distributional semantic models are a fundamental component in many natural language processing (NLP) architectures.
Approach: They propose a multi-lingual word embedding/distributional semantics framework which supports creation, use and evaluation of word embedded models.
Outcome: The proposed tool supports the creation, use and evaluation of word embedding models.
Disentangling Generative Factors in Natural Language with Discrete Variational Autoencoders (2021.findings-emnlp)

Copied to clipboard

Challenge: Disentangled representation learning aims to provide an interpretable representation of latent features and a framework for controlling the change of specific features.
Approach: They propose a Variational Autoencoder based method which models language features as discrete variables and encourages independence between variables for learning disentangled representations.
Outcome: The proposed model outperforms baselines on several qualitative and quantitative benchmarks and on a text style transfer downstream application.
A Survey on Open Information Extraction (C18-1)

Copied to clipboard

Challenge: Existing approaches to open information extraction (Open IE) focus on narrow, well-defined requests over a predefined set of target relations on small, homogeneous corpora.
Approach: They propose to use unsupervised methods to extract all types of relations found in text . they propose to implement a system that can be automated to detect possible relations .
Outcome: The proposed approaches have been compared with existing methods and are based on the results of a literature review.
Explainable Inference Over Grounding-Abstract Chains for Science Questions (2021.findings-acl)

Copied to clipboard

Challenge: Existing inference models for science questions are black-box by nature, lacking explanations for their predictions.
Approach: They propose an explainable inference approach for science questions by reasoning on grounding and abstract inference chains.
Outcome: The proposed model generates plausible explanations for science questions using a weighted graph of relevant facts and a Bayesian Optimisation formalism.
Natural Language Premise Selection: Finding Supporting Statements for Mathematical Text (2020.lrec-1)

Copied to clipboard

Challenge: Existing approaches to understand mathematical discourse are limited by the complexity of word and symbol interactions.
Approach: They propose a task to retrieve supporting definitions and supporting propositions from a dataset that can be used to evaluate different approaches for the task.
Outcome: The proposed task is based on a dataset that can be used to evaluate different approaches for the natural premise selection task.
Neuro-Symbolic Natural Language Processing (2025.emnlp-tutorials)

Copied to clipboard

Challenge: Large Language Models (LLMs) have limitations in terms of safe and controlled reasoning, interpretability and adaptability . this tutorial aims to bridge the gap between the practical performance of LLMs and the principled modelling of language and inference of formal methods.
Approach: This tutorial aims to bridge the gap between the practical performance of Large Language Models and the principled modelling of language and inference of formal methods.
Outcome: This tutorial aims to bridge the gap between the performance of LLMs and the principled modelling of language and inference of formal methods.
The SSIX Corpora: Three Gold Standard Corpora for Sentiment Analysis in English, Spanish and German Financial Microblogs (L18-1)

Copied to clipboard

Challenge: SSIX corpora provide annotated data for supervised learning methods . polarity annotation is performed on two financial microblog platforms .
Approach: They propose three SSIX corpora for sentiment analysis which provide annotated data for supervised learning methods.
Outcome: The proposed corpora are in English, German and Spanish.
Identifying Supporting Facts for Multi-hop Question Answering with Document Graph Networks (D19-53)

Copied to clipboard

Challenge: Recent advances in reading comprehension have resulted in models that surpass human performance when the answer is contained in a single, continuous passage of text.
Approach: They propose a document-structured message passing architecture for the identification of supporting facts over a graph-structure based representation of text.
Outcome: The proposed model outperforms a baseline reading comprehension test on raw text and shows that it is relevant for multi-hop reasoning.
STAR: Cross-modal [STA]tement [R]epresentation for selecting relevant mathematical premises (2021.eacl-main)

Copied to clipboard

Challenge: Existing representations of mathematical statements in natural language are ineffective . STAR model uses cross-modal attention to represent mathematical text .
Approach: They propose a model that uses cross-modal attention to represent mathematical text . it uses conjectures written in both natural and mathematical language to recommend premises .
Outcome: The proposed model outperforms baseline models that do not distinguish between natural and mathematical elements and achieves better performance than state-of-the-art models.
Interventional Probing in High Dimensions: An NLI Case Study (2023.findings-eacl)

Copied to clipboard

Challenge: Probing strategies have been shown to detect the presence of various linguistic features inlarge language models; in particular, semantic features intermediate to the “natural logic”fragment of the NLI.
Approach: They propose to use amnesic probing and mnestic probing to investigate the effect of these semantic fea-tures on NLI classification by examining the effects of a mnemonic probing variation on the model.
Outcome: The proposed methods have been shown to detect features intermediate to the “natural logic”fragment of the Natural Language Inferencetask (NLI).
Case-Based Abductive Natural Language Inference (2022.coling-1)

Copied to clipboard

Challenge: Recent approaches for multi-hop inference construct explanations considering each test case in isolation, but semantic drift causes wrong conclusions.
Approach: They propose an abductive framework for multi-hop NLI exploring the retrieve-reuse-refine paradigm in Case-Based Reasoning.
Outcome: The proposed model can be integrated with sparse and dense pre-trained encoders to improve multi-hop inference, or adopted as an evidence retriever for Transformers.
A Survey of Text Games for Reinforcement Learning Informed by Natural Language (2022.tacl-1)

Copied to clipboard

Challenge: Interactive Fiction Games (text games) are a problem type that require natural language to solve complex tasks.
Approach: They propose to use interactive fiction games as a testing environment to test the new Reinforcement Learning solutions using natural language.
Outcome: The proposed solutions are based on the proposed interactive fiction games and the generated environments.
Graphene: Semantically-Linked Propositions in Open Information Extraction (C18-1)

Copied to clipboard

Challenge: Existing Open IE systems focus on identifying and extracting relations of interest, but this manual labor scales linearly with the number of target relations.
Approach: They propose an Open Information Extraction approach that uses a two-layered transformation stage and rhetorical relation identification to transform sentences into syntactically sound sentences.
Outcome: The proposed approach outperforms state-of-the-art Open IE systems in the construction of correct n-ary predicate-argument structures.

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations