Publications

Using Interactive Feedback to Improve the Accuracy and Explainability of Question Answering Systems Post-Deployment
Zichao Li
Prakhar Sharma
Xing Han Lu
Jackie Cheung
Siva Reddy
Most research on question answering focuses on the pre-deployment stage; i.e., building an accurate model for deployment.In this paper, we a… (voir plus)sk the question: Can we improve QA systems further post-deployment based on user interactions? We focus on two kinds of improvements: 1) improving the QA system’s performance itself, and 2) providing the model with the ability to explain the correctness or incorrectness of an answer.We collect a retrieval-based QA dataset, FeedbackQA, which contains interactive feedback from users. We collect this dataset by deploying a base QA system to crowdworkers who then engage with the system and provide feedback on the quality of its answers.The feedback contains both structured ratings and unstructured natural language explanations.We train a neural model with this feedback data that can generate explanations and re-score answer candidates. We show that feedback data not only improves the accuracy of the deployed QA system but also other stronger non-deployed systems. The generated explanations also help users make informed decisions about the correctness of answers.
2022-04-06
ArXiv (preprint)
doi.org
arxiv.org
Comparison of multicenter MRI protocols for visualizing the spinal cord gray matter
Julien Cohen-Adad
Eva Alonso‐Ortiz
Stephanie Alley
M. M. Laganá
Francesca Baglio
S. Vannesjo
Haleh Karbasforoushan
Maryam Seif
A. Seifert
Junqian Xu
Joo-won Kim
René Labounek
Lubomír Vojtíšek
Marek Dostál
Jan Valosek
Rebecca Sara Samson
Francesco Grussu
Marco Battiston
C. G. Gandini Wheeler-Kingshott
Marios C. Yiannakas … (voir 4 de plus)
Guillaume Gilbert
Torben Schneider
Brian Johnson
Ferran Prados
Spinal cord gray‐matter imaging is valuable for a number of applications, but remains challenging. The purpose of this work was to compare… (voir plus) various MRI protocols at 1.5 T, 3 T, and 7 T for visualizing the gray matter.
2022-04-05
Magnetic Resonance in Medicine (published)
doi.org
Comparison of multicenter MRI protocols for visualizing the spinal cord gray matter
Julien Cohen-Adad
Eva Alonso‐Ortiz
Stephanie Alley
Maria Marcella Lagana
Francesca Baglio
Signe Johanna Vannesjo
Haleh Karbasforoushan
Maryam Seif
Alan C. Seifert
Junqian Xu
Joo‐Won Kim
René Labounek
Lubomír Vojtíšek
Marek Dostál
Jan Valosek
Rebecca S. Samson
Francesco Grussu
Marco Battiston
Claudia A. M. Gandini Wheeler-Kingshott
Marios C. Yiannakas … (voir 4 de plus)
Guillaume Gilbert
Torben Schneider
Brian Johnson
Ferran Prados
Spinal cord gray‐matter imaging is valuable for a number of applications, but remains challenging. The purpose of this work was to compare… (voir plus) various MRI protocols at 1.5 T, 3 T, and 7 T for visualizing the gray matter.
2022-04-05
Magnetic Resonance in Medicine (publié)
doi.org
Comparison of multicenter MRI protocols for visualizing the spinal cord gray matter
Julien Cohen-Adad
Eva Alonso‐Ortiz
Stephanie Alley
Maria Marcella Lagana
Francesca Baglio
Signe Johanna Vannesjo
Haleh Karbasforoushan
Maryam Seif
Alan C. Seifert
Junqian Xu
Joo‐Won Kim
René Labounek
Lubomír Vojtíšek
Marek Dostál
Jan Valosek
Rebecca S. Samson
Francesco Grussu
Marco Battiston
Claudia A. M. Gandini Wheeler-Kingshott
Marios C. Yiannakas … (voir 4 de plus)
Guillaume Gilbert
Torben Schneider
Brian Johnson
Ferran Prados
Spinal cord gray‐matter imaging is valuable for a number of applications, but remains challenging. The purpose of this work was to compare… (voir plus) various MRI protocols at 1.5 T, 3 T, and 7 T for visualizing the gray matter.
2022-04-05
Magnetic Resonance in Medicine (publié)
doi.org
E VALUATING G ENERALIZATION IN GF LOW N ETS FOR M OLECULE D ESIGN
Andrei Cristian Nica
Moksh J. Jain
Emmanuel Bengio
Cheng-Hao Liu
Maksym Korablyov
Michael M. Bronstein
Yoshua Bengio
Deep learning bears promise for drug discovery problems such as de novo molecular design. Generating data to train such models is a costly a… (voir plus)nd time-consuming process, given the need for wet-lab experiments or expensive simulations. This problem is compounded by the notorious data-hungriness of machine learning algorithms. In small molecule generation the recently proposed GFlowNet method has shown good performance in generating diverse high-scoring candidates, and has the interesting advantage of being an off-policy offline method. Finding an appropriate generalization evaluation metric for such models, one predictive of the desired search performance (i.e. finding high-scoring diverse candidates), will help guide online data collection for such an algorithm. In this work, we develop techniques for evaluating GFlowNet performance on a test set, and identify the most promising metric for predicting generalization. We present empirical results on several small-molecule design tasks in drug discovery, for several GFlowNet training setups, and we find a metric strongly correlated with diverse high-scoring batch generation. This metric should be used to identify the best generative model from which to sample batches of molecules to be evaluated.
2022-04-05
ICLR.cc/2022/Workshop/MLDD (poster)
openreview.net
openreview.net
Naming Autism in the Right Context.
Andres Roman-Urrestarazu
Guillaume Dumas
Varun Warrier
2022-04-04
JAMA pediatrics (published)
doi.org
Tracking and predicting COVID-19 radiological trajectory on chest X-rays using deep learning
Daniel Gourdeau
Olivier Potvin
Patrick Archambault
Carl Chartrand‐lefebvre
Louis Dieumegarde
Reza Forghani
Christian Gagné
Alexandre Hains
David Hornstein
Huy Khiem Le
Simon Lemieux
Marie‐hélène Lévesque
Diego R. Martin
Lorne Rosenbloom
An Tang
Fabrizio Vecchio
Issac Y Yang
N. Duchesne
Simon Duchesne
2022-04-04
Scientific Reports (published)
doi.org
Tracking and predicting COVID-19 radiological trajectory on chest X-rays using deep learning
Daniel Gourdeau
Olivier Potvin
Patrick Archambault
Carl Chartrand-Lefebvre
Louis Dieumegarde
Reza Forghani
Christian Gagné
Alexandre Hains
David Hornstein
Huy Le
Simon Lemieux
Marie-Hélène Lévesque
Diego Martin
Lorne Rosenbloom
An Tang
Fabrizio Vecchio
Issac Yang
Nathalie Duchesne
Simon Duchesne
Radiological findings on chest X-ray (CXR) have shown to be essential for the proper management of COVID-19 patients as the maximum severity… (voir plus) over the course of the disease is closely linked to the outcome. As such, evaluation of future severity from current CXR would be highly desirable. We trained a repurposed deep learning algorithm on the CheXnet open dataset (224,316 chest X-ray images of 65,240 unique patients) to extract features that mapped to radiological labels. We collected CXRs of COVID-19-positive patients from an open-source dataset (COVID-19 image data collection) and from a multi-institutional local ICU dataset. The data was grouped into pairs of sequential CXRs and were categorized into three categories: ‘Worse’, ‘Stable’, or ‘Improved’ on the basis of radiological evolution ascertained from images and reports. Classical machine-learning algorithms were trained on the deep learning extracted features to perform immediate severity evaluation and prediction of future radiological trajectory. Receiver operating characteristic analyses and Mann-Whitney tests were performed. Deep learning predictions between “Worse” and “Improved” outcome categories and for severity stratification were significantly different for three radiological signs and one diagnostic (‘Consolidation’, ‘Lung Lesion’, ‘Pleural effusion’ and ‘Pneumonia’; all P 0.05). Features from the first CXR of each pair could correctly predict the outcome category between ‘Worse’ and ‘Improved’ cases with a 0.81 (0.74–0.83 95% CI) AUC in the open-access dataset and with a 0.66 (0.67–0.64 95% CI) AUC in the ICU dataset. Features extracted from the CXR could predict disease severity with a 52.3% accuracy in a 4-way classification. Severity evaluation trained on the COVID-19 image data collection had good out-of-distribution generalization when testing on the local dataset, with 81.6% of intubated ICU patients being classified as critically ill, and the predicted severity was correlated with the clinical outcome with a 0.639 AUC. CXR deep learning features show promise for classifying disease severity and trajectory. Once validated in studies incorporating clinical data and with larger sample sizes, this information may be considered to inform triage decisions.
2022-04-04
Scientific Reports (publié)
doi.org
Tracking and predicting COVID-19 radiological trajectory on chest X-rays using deep learning
Daniel Gourdeau
Olivier Potvin
Patrick Archambault
Carl Chartrand-Lefebvre
Louis Dieumegarde
Reza Forghani
Christian Gagné
Alexandre Hains
David Hornstein
Huy Le
Simon Lemieux
Marie-Hélène Lévesque
Diego Martin
Lorne Rosenbloom
An Tang
Fabrizio Vecchio
Issac Yang
Nathalie Duchesne
Simon Duchesne
2022-04-04
Scientific Reports (publié)
doi.org
TRACKING AND PREDICTING COVID-19 RADIOLOGICAL TRAJECTORY USING DEEP LEARNING ON CHEST X-RAYS: INITIAL ACCURACY TESTING
Simon Duchesne
Olivier Potvin
Daniel Gourdeau
Patrick Archambault
Carl Chartrand-Lefebvre
Louis Dieumegarde
Reza Forghani
Christian Gagné
Alexandre Hains
David Hornstein
Huy Le
Simon Lemieux
Marie-Hélène Lévesque
Diego Martin
Lorne Rosenbloom
An Tang
Fabrizio Vecchio
Issac Yang
Nathalie Duchesne
2022-04-04
Scientific Reports (publié)
doi.org
Why Exposure Bias Matters: An Imitation Learning Perspective of Error Accumulation in Language Generation
Kushal Arora
Layla El Asri
Hareesh Bahuleyan
Jackie Cheung
Current language generation models suffer from issues such as repetition, incoherence, and hallucinations. An often-repeated hypothesis for … (voir plus)this brittleness of generation models is that it is caused by the training and the generation procedure mismatch, also referred to as exposure bias. In this paper, we verify this hypothesis by analyzing exposure bias from an imitation learning perspective. We show that exposure bias leads to an accumulation of errors during generation, analyze why perplexity fails to capture this accumulation of errors, and empirically show that this accumulation results in poor generation quality.
2022-04-03
ArXiv (prépublication)
doi.org
arxiv.org
Matching Feature Sets for Few-Shot Image Classification
Arman Afrasiyabi
Hugo Larochelle
Jean‐François Lalonde
Christian Gagné
In image classification, it is common practice to train deep networks to extract a single feature vector per input image. Few-shot classific… (voir plus)ation methods also mostly follow this trend. In this work, we depart from this established direction and instead propose to extract sets of feature vectors for each image. We argue that a set-based representation intrinsically builds a richer representation of images from the base classes, which can subsequently better transfer to the few-shot classes. To do so, we propose to adapt existing feature extractors to instead produce sets of feature vectors from images. Our approach, dubbed SetFeat, embeds shallow self-attention mechanisms inside existing encoder architectures. The attention modules are lightweight, and as such our method results in encoders that have approximately the same number of parameters as their original versions. During training and inference, a set-to-set matching metric is used to perform image classification. The effectiveness of our proposed architecture and metrics is demonstrated via thorough experiments on standard few-shot datasets-namely miniImageNet, tieredImageNet, and CUB-in both the 1- and 5-shot scenarios. In all cases but one, our method outperforms the state-of-the-art.
2022-04-02
ArXiv (preprint)
doi.org
arxiv.org