Portrait de Siva Reddy

Siva Reddy

Membre académique principal
Chaire en IA Canada-CIFAR
Professeur adjoint, McGill University
Sujets de recherche
Apprentissage de représentations
Apprentissage profond
Raisonnement
Traitement du langage naturel

Biographie

Siva Reddy est professeur adjoint en informatique et linguistique à l’Université McGill. Ses travaux portent sur les algorithmes qui permettent aux ordinateurs de comprendre et de traiter les langues humaines. Il a fait ses études postdoctorales avec le Stanford NLP Group. Son expertise inclut la construction de symboliques linguistiques et induites et de modèles d’apprentissage profond pour le langage.

Étudiants actuels

Postdoctorat - McGill
Superviseur⋅e principal⋅e :
Stagiaire de recherche - McGill
Maîtrise recherche - McGill
Co-superviseur⋅e :
Doctorat - McGill
Co-superviseur⋅e :
Doctorat - McGill
Superviseur⋅e principal⋅e :
Doctorat - McGill
Co-superviseur⋅e :
Doctorat - McGill
Co-superviseur⋅e :
Maîtrise recherche - McGill
Co-superviseur⋅e :
Doctorat - McGill
Superviseur⋅e principal⋅e :
Stagiaire de recherche - McGill
Collaborateur·rice de recherche - University of Edinburgh
Collaborateur·rice de recherche - N/A
Collaborateur·rice de recherche - McGill
Collaborateur·rice de recherche - McGill

Publications

Using Interactive Feedback to Improve the Accuracy and Explainability of Question Answering Systems Post-Deployment
TopiOCQA: Open-domain Conversational Question Answering with Topic Switching
Shehzaad Dhuliawala
Kaheer Suleman
Combining Modular Skills in Multitask Learning
Edoardo M. Ponti
A modular design encourages neural models to disentangle and recombine different facets of knowledge to generalise more systematically to ne… (voir plus)w tasks. In this work, we assume that each task is associated with a subset of latent discrete skills from a (potentially small) inventory. In turn, skills correspond to parameter-efficient (sparse / low-rank) model parameterisations. By jointly learning these and a task-skill allocation matrix, the network for each task is instantiated as the average of the parameters of active skills. To favour non-trivial soft partitions of skills across tasks, we experiment with a series of inductive biases, such as an Indian Buffet Process prior and a two-speed learning rate. We evaluate our latent-skill model on two main settings: 1) multitask reinforcement learning for grounded instruction following on 8 levels of the BabyAI platform; and 2) few-shot adaptation of pre-trained text-to-text generative models on CrossFit, a benchmark comprising 160 NLP tasks. We find that the modular design of a network significantly increases sample efficiency in reinforcement learning and few-shot generalisation in supervised learning, compared to baselines with fully shared, task-specific, or conditionally generated parameters where knowledge is entangled across tasks. In addition, we show how discrete skills help interpretability, as they yield an explicit hierarchy of tasks.
Image Retrieval from Contextual Descriptions
Vibhav Vineet
Edoardo Ponti
The ability to integrate context, including perceptual and temporal cues, plays a pivotal role in grounding the meaning of a linguistic utte… (voir plus)rance. In order to measure to what extent current vision-and-language models master this ability, we devise a new multimodal challenge, Image Retrieval from Contextual Descriptions (ImageCoDe). In particular, models are tasked with retrieving the correct image from a set of 10 minimally contrastive candidates based on a contextual description.As such, each description contains only the details that help distinguish between images.Because of this, descriptions tend to be complex in terms of syntax and discourse and require drawing pragmatic inferences. Images are sourced from both static pictures and video frames.We benchmark several state-of-the-art models, including both cross-encoders such as ViLBERT and bi-encoders such as CLIP, on ImageCoDe.Our results reveal that these models dramatically lag behind human performance: the best variant achieves an accuracy of 20.9 on video frames and 59.4 on static pictures, compared with 90.8 in humans.Furthermore, we experiment with new model variants that are better equipped to incorporate visual and temporal context into their representations, which achieve modest gains. Our hope is that ImageCoDE will foster progress in grounded language understanding by encouraging models to focus on fine-grained visual differences.
On the Origin of Hallucinations in Conversational Models: Is it the Datasets or the Models?
Nouha Dziri
Mo Yu
Osmar Zaiane
Knowledge-grounded conversational models are known to suffer from producing factually invalid statements, a phenomenon commonly called hallu… (voir plus)cination. In this work, we investigate the underlying causes of this phenomenon: is hallucination due to the training data, or to the models? We conduct a comprehensive human study on both existing knowledge-grounded conversational benchmarks and several state-of-the-art models. Our study reveals that the standard benchmarks consist of >60% hallucinated responses, leading to models that not only hallucinate but even amplify hallucinations. Our findings raise important questions on the quality of existing datasets and models trained using them. We make our annotations publicly available for future research.
Understanding by Understanding Not: Modeling Negation in Language Models
Negation is a core construction in natural language. Despite being very successful on many tasks, state-of-the-art pre-trained language mode… (voir plus)ls often handle negation incorrectly. To improve language models in this regard, we propose to augment the language modeling objective with an unlikelihood objective that is based on negated generic sentences from a raw text corpus. By training BERT with the resulting combined objective we reduce the mean top~1 error rate to 4% on the negated LAMA dataset. We also see some improvements on the negated NLI benchmarks.
End-to-End Training of Multi-Document Reader and Retriever for Open-Domain Question Answering
Devendra Singh Sachan
William Hamilton
Chris Dyer
Dani Yogatama
We present an end-to-end differentiable training method for retrieval-augmented open-domain question answering systems that combine informat… (voir plus)ion from multiple retrieved documents when generating answers. We model retrieval decisions as latent variables over sets of relevant documents. Since marginalizing over sets of retrieved documents is computationally hard, we approximate this using an expectation-maximization algorithm. We iteratively estimate the value of our latent variable (the set of relevant documents for a given question) and then use this estimate to update the retriever and reader parameters. We hypothesize that such end-to-end training allows training signals to flow to the reader and then to the retriever better than staged-wise training. This results in a retriever that is able to select more relevant documents for a question and a reader that is trained on more accurate documents to generate an answer. Experiments on three benchmark datasets demonstrate that our proposed method outperforms all existing approaches of comparable size by 2-3% absolute exact match points, achieving new state-of-the-art results. Our results also demonstrate the feasibility of learning to retrieve to improve answer generation without explicit supervision of retrieval decisions.
Back-Training excels Self-Training at Unsupervised Domain Adaptation of Question Generation and Passage Retrieval
Robert Belfer
Iulian V. Serban
In this work, we introduce back-training, an alternative to self-training for unsupervised domain adaptation (UDA). While self-training gene… (voir plus)rates synthetic training data where natural inputs are aligned with noisy outputs, back-training results in natural outputs aligned with noisy inputs. This significantly reduces the gap between target domain and synthetic data distribution, and reduces model overfitting to source domain. We run UDA experiments on question generation and passage retrieval from the Natural Questions domain to machine learning and biomedical domains. We find that back-training vastly outperforms self-training by a mean improvement of 7.8 BLEU-4 points on generation, and 17.6% top-20 retrieval accuracy across both domains. We further propose consistency filters to remove low-quality synthetic data before training. We also release a new domain-adaptation dataset - MLQuestions containing 35K unaligned questions, 50K unaligned passages, and 3K aligned question-passage pairs.
Visually Grounded Reasoning across Languages and Cultures
Fangyu Liu
Emanuele Bugliarello
Edoardo Ponti
Nigel Collier
The design of widespread vision-and-language datasets and pre-trained encoders directly adopts, or draws inspiration from, the concepts and … (voir plus)images of ImageNet. While one can hardly overestimate how much this benchmark contributed to progress in computer vision, it is mostly derived from lexical databases and image queries in English, resulting in source material with a North American or Western European bias. Therefore, we devise a new protocol to construct an ImageNet-style hierarchy representative of more languages and cultures. In particular, we let the selection of both concepts and images be entirely driven by native speakers, rather than scraping them automatically. Specifically, we focus on a typologically diverse set of languages, namely, Indonesian, Mandarin Chinese, Swahili, Tamil, and Turkish. On top of the concepts and images obtained through this new protocol, we create a multilingual dataset for Multicultural Reasoning over Vision and Language (MaRVL) by eliciting statements from native speaker annotators about pairs of images. The task consists of discriminating whether each grounded statement is true or false. We establish a series of baselines using state-of-the-art models and find that their cross-lingual transfer performance lags dramatically behind supervised performance in English. These results invite us to reassess the robustness and accuracy of current state-of-the-art models beyond a narrow domain, but also open up new exciting challenges for the development of truly multilingual and multicultural systems.
Abg-CoQA: Clarifying Ambiguity in Conversational Question Answering
Meiqi Guo
Mingda Zhang
Malihe Alikhani
Effective communication is about the dissemination of properly worded meaningful ideas/messages that are comprehensible to both sen… (voir plus)der and receiver and which ultimately can attract the desired response or feedback. For machines to engage in a conversation, it is therefore essential to enable them to clarify ambiguity and achieve a common ground. We introduce Abg-CoQA, a novel dataset for clarifying ambiguity in Conversational Question Answering systems. Our dataset contains 9k questions with answers where 1k questions are ambiguous, obtained from 4k text passages from five diverse domains. For ambiguous questions, a clarification conversational turn is collected. We evaluate strong language generation models and conversational question answering models on Abg-CoQA. The best-performing system achieves a BLEU-1 score of 12.9% on generating clarification question, which is 27.9 points behind human performance (40.8%); and a F1 score of 40.1% on question answering after clarification, which is 35.1 points behind human performance (75.2%), indicating there is ample room for improvement.
Minimax and Neyman-Pearson Meta-Learning for Outlier Languages
Disha Shrivastava
Anders Søgaard
Model-agnostic meta-learning (MAML) has been recently put forth as a strategy to learn resource-poor languages in a sample-efficient fashion… (voir plus). Nevertheless, the properties of these languages are often not well represented by those available during training. Hence, we argue that the i.i.d. assumption ingrained in MAML makes it ill-suited for cross-lingual NLP. In fact, under a decision-theoretic framework, MAML can be interpreted as minimising the expected risk across training languages (with a uniform prior), which is known as Bayes criterion. To increase its robustness to outlier languages, we create two variants of MAML based on alternative criteria: Minimax MAML reduces the maximum risk across languages, while Neyman-Pearson MAML constrains the risk in each language to a maximum threshold. Both criteria constitute fully differentiable two-player games. In light of this, we propose a new adaptive optimiser solving for a local approximation to their Nash equilibrium. We evaluate both model variants on two popular NLP tasks, part-of-speech tagging and question answering. We report gains for their average and minimum performance across low-resource languages in zero- and few-shot settings, compared to joint multi-source transfer and vanilla MAML.
StereoSet: Measuring stereotypical bias in pretrained language models
Moin Nadeem
Anna Bethke
A stereotype is an over-generalized belief about a particular group of people, e.g., Asians are good at math or African Americans are athlet… (voir plus)ic. Such beliefs (biases) are known to hurt target groups. Since pretrained language models are trained on large real-world data, they are known to capture stereotypical biases. It is important to quantify to what extent these biases are present in them. Although this is a rapidly growing area of research, existing literature lacks in two important aspects: 1) they mainly evaluate bias of pretrained language models on a small set of artificial sentences, even though these models are trained on natural data 2) current evaluations focus on measuring bias without considering the language modeling ability of a model, which could lead to misleading trust on a model even if it is a poor language model. We address both these problems. We present StereoSet, a large-scale natural English dataset to measure stereotypical biases in four domains: gender, profession, race, and religion. We contrast both stereotypical bias and language modeling ability of popular models like BERT, GPT-2, RoBERTa, and XLnet. We show that these models exhibit strong stereotypical biases. Our data and code are available at https://stereoset.mit.edu.