Papers by Edoardo Maria Ponti
Probing Cross-Lingual Lexical Knowledge from Multilingual Sentence Encoders (2023.eacl-main)
Copied to clipboard
| Challenge: | Pretrained multilingual language models (LMs) can be 'rewired' into effective multilingual sentence encoders (SEs) however, it remains unclear how to best leverage them to represent sub-sentence lexical items in cross-lingual lexicals. |
| Approach: | They propose a method for exposing cross-lingual lexical knowledge by additional fine-tuning through inexpensive contrastive learning that requires only a small amount of word translation pairs. |
| Outcome: | The proposed method exposes cross-lingual lexical knowledge by additional fine-tuning through inexpensive contrastive learning that requires only a small amount of word translation pairs. |
Isomorphic Transfer of Syntactic Structures in Cross-Lingual NLP (P18-1)
Copied to clipboard
| Challenge: | Using frameworks such as Universal Dependencies (UD) to transfer knowledge between languages can be challenging because of variation in syntactic structures. |
| Approach: | They propose a typologically driven method which reduces anisomorphism in UD treebanks by considering both morphological and structural properties. |
| Outcome: | The proposed method is effective for machine translation and cross-lingual sentence similarity. |
On the Relation between Linguistic Typology and (Limitations of) Multilingual Language Modeling (D18-1)
Copied to clipboard
| Challenge: | a key challenge in cross-lingual NLP is developing general language-independent architectures that are equally applicable to any language. |
| Approach: | They propose to use a full-vocabulary setup to test the performance of language modeling (LM) on 50 typologically diverse languages. |
| Outcome: | The proposed language modeling task is based on a full vocabulary setup focused on word-level prediction on 50 typologically diverse languages. |
XCOPA: A Multilingual Dataset for Causal Commonsense Reasoning (2020.emnlp-main)
Copied to clipboard
| Challenge: | XCOPA dataset provides a typologically diverse dataset for commonsense reasoning in 11 languages . current methods for evaluating commonsensible reasoning in resource-poor languages are weak compared to translation-based transfer. |
| Approach: | They propose a typologically diverse multilingual dataset for causal commonsense reasoning in 11 languages. |
| Outcome: | The proposed model performs better than current methods on a resource-poor dataset compared to translation-based transfer in the 11 languages studied . |
Visually Grounded Reasoning across Languages and Cultures (2021.emnlp-main)
Copied to clipboard
| Challenge: | a new protocol allows for a multilingual hierarchy of concepts and images based on native speakers . the results suggest that the current models are not robust enough to handle multilingual data . |
| Approach: | They propose a protocol to construct an ImageNet-style hierarchy representative of more languages and cultures. |
| Outcome: | The proposed protocol lets the selection of concepts and images be entirely driven by native speakers, rather than scraping them automatically. |
Verb Knowledge Injection for Multilingual Event Processing (2021.acl-long)
Copied to clipboard
| Challenge: | Recent studies have shown that pretrainers implicitly extract a non-negligible amount of linguistic knowledge from text corpora in an unsupervised fashion. |
| Approach: | They propose to inject explicit verb knowledge into dedicated adapter modules to complement the linguistic knowledge obtained during LM-pretraining. |
| Outcome: | The proposed model improves in English event extraction tasks, while injecting verb knowledge improves other languages. |
Internal and external pressures on language emergence: least effort, object constancy and frequency (2020.findings-emnlp)
Copied to clipboard
| Challenge: | Existing studies show that the emergent languages rarely display salient features inherent to natural languages, such as compositionality of meaning and generalisation to novel objects. |
| Approach: | They propose to formalise the principle of least effort through an auxiliary objective and explore several game variants inspired by the principle 'object constancy' they find that the proposed sources of pressure result in emerging languages with less redundancy, more focus on high-level conceptual information, and better abilities of generalisation. |
| Outcome: | The proposed sources of pressure result in emerging languages with less redundancy, more focus on high-level conceptual information, and better abilities of generalisation. |
Combining Parameter-efficient Modules for Task-level Generalisation (2023.eacl-main)
Copied to clipboard
| Challenge: | A modular design encourages neural models to disentangle and recombine different facets of knowledge to generalise more systematically to new tasks. |
| Approach: | They propose a modular neural network where a subset of latent skills is associated with a parameter-efficient model adapter. |
| Outcome: | The proposed model improves sample efficiency and few-shot generalisation in supervised learning compared to baselines. |
AM2iCo: Evaluating Word Meaning in Context across Low-Resource Languages with Adversarial Examples (2021.emnlp-main)
Copied to clipboard
| Challenge: | Existing multilingual evaluation datasets that evaluate lexical semantics "in-context" have various limitations, including limited coverage of high-resource languages and superficial cues. |
| Approach: | They propose to use a set of pretrained language models to evaluate lexical semantics in context. |
| Outcome: | The proposed set shows that current models lag behind human performance in interpreting word meaning in cross-lingual contexts. |
Cross-lingual Semantic Specialization via Lexical Relation Induction (D19-1)
Copied to clipboard
| Challenge: | Semantic specialization is not available in many languages because of their incomplete or non-existent structure. |
| Approach: | They propose a method that transfers specialization from a resource-rich source language to virtually any target language. |
| Outcome: | The proposed method performs lexical simplification, dialog state tracking, and textual similarity tasks in 5 languages. |
MAD-G: Multilingual Adapter Generation for Efficient Cross-Lingual Transfer (2021.findings-emnlp)
Copied to clipboard
Alan Ansell, Edoardo Maria Ponti, Jonas Pfeiffer, Sebastian Ruder, Goran Glavaš, Ivan Vulić, Anna Korhonen
| Challenge: | Massively multilingual transformers (MMTs) have benefited from additional training of language-specific adapters, but this approach is not viable for the vast majority of languages due to limitations in their corpus size or compute budgets. |
| Approach: | They propose a multilingual ADapter generation approach which contextually generates language adapters from language representations based on typological features. |
| Outcome: | The proposed method improves cross-lingual transfer performance on part-of-speech tagging, dependency parsing, and named entity recognition tasks while remaining cost-effective. |
Specializing Unsupervised Pretraining Models for Word-Level Semantic Similarity (2020.coling-main)
Copied to clipboard
| Challenge: | Unsupervised pretraining models encode only distributional knowledge encoded in text corpora, incorporated through language modeling objectives. |
| Approach: | They generalize a standard BERT model to a multi-task learning setting and integrate discrete knowledge on word-level semantic similarity into pretraining. |
| Outcome: | The proposed model outperforms the lexically blind “vanilla” model on several language understanding tasks. |
Distilling Efficient Language-Specific Models for Cross-Lingual Transfer (2023.findings-acl)
Copied to clipboard
| Challenge: | Massively multilingual Transformers (MMTs) are widely used for cross-lingual transfer learning. |
| Approach: | They propose to extract compressed, language-specific models from MMTs which retain the capacity of the original MMT for cross-lingual transfer. |
| Outcome: | The proposed model outperforms models trained from scratch in zero-shot cross-lingual transfer across benchmarks. |
Adversarial Propagation and Zero-Shot Cross-Lingual Transfer of Word Vector Specialization (D18-1)
Copied to clipboard
| Challenge: | Semantic specialization is a process of fine-tuning pre-trained distributional word vectors using external lexical knowledge to accentuate a particular semantic relation in the specialized vector space. |
| Approach: | They propose a method for specializing distributional word vectors using external lexical knowledge. |
| Outcome: | The proposed method improves on word similarity, dialog state tracking, and lexical simplification across three languages and on three tasks. |
Efficient Transformers with Dynamic Token Pooling (2023.acl-long)
Copied to clipboard
| Challenge: | Hourglass Transformers is a computationally efficient model that can be used to reduce the sequence length in the intermediate layers. |
| Approach: | They propose a dynamic-pooling mechanism which predicts segment boundaries in an autoregressive fashion. |
| Outcome: | The proposed model is faster and more accurate than vanilla Transformers and fixed-length pooling within the same computational budget. |
Probing Pretrained Language Models for Lexical Semantics (2020.emnlp-main)
Copied to clipboard
| Challenge: | Existing studies have focused on morphosyntactic, semantic, and world knowledge, but it remains unclear to what extent LMs derive lexical type-level knowledge from words in context. |
| Approach: | They propose to use multilingual and monolingual LMs to extract lexical type-level knowledge from words in context. |
| Outcome: | The proposed models perform well across six typologically diverse languages and five lexical tasks. |
Towards Zero-shot Language Modeling (D19-1)
Copied to clipboard
| Challenge: | a number of natural questions have been asked about the inductive biases of neural networks on core NLP tasks. |
| Approach: | They construct an informative prior for held-out languages on a task of character-level, open-vocabulary language modelling. |
| Outcome: | The proposed model outperforms baseline models with an uninformative prior in both zero-shot and few-shot settings, showing that it is imbued with universal linguistic knowledge. |
Minimax and Neyman–Pearson Meta-Learning for Outlier Languages (2021.findings-acl)
Copied to clipboard
| Challenge: | Model-agnostic meta-learning (MAML) is a strategy to learn resource-poor languages in a sample-efficient fashion. |
| Approach: | They propose a model-agnostic meta-learning strategy that minimizes the expected risk across languages with a uniform prior . they propose 'minimax' and 'neyman-pearson' models that constrain the risk in each language to a maximum threshold. |
| Outcome: | The proposed model reduces the maximum risk across languages while constraining the risk in each language to a maximum threshold. |
LexFit: Lexical Fine-Tuning of Pretrained Language Models (2021.acl-long)
Copied to clipboard
| Challenge: | Transformer-based language models implicitly store a wealth of lexical semantic knowledge, but it is non-trivial to extract that knowledge effectively from their parameters. |
| Approach: | They propose to expose and enrich lexical knowledge from transformer-based language models to serve as effective decontextualized word encoders even when fed input words "in isolation" |
| Outcome: | The proposed model outperforms standard static WEs and vanilla LMs in lexical tasks over four established tasks in 8 languages. |
Emergent Communication Pretraining for Few-Shot Machine Translation (2020.coling-main)
Copied to clipboard
| Challenge: | state-of-the-art models that rely on multilingual pretrained encoders achieve sample efficiency in downstream applications, but lack abundant amounts of unlabelled text. |
| Approach: | They propose a method to pretrain neural networks via emergent communication from referential games by grounding communication on images as a crude approximation of real-world environments. |
| Outcome: | The proposed method significantly improves machine translation in few-shot settings and provides an evaluation protocol to probe the properties of emergent languages ex vitro. |