Publications

"Your child needs surgery": A survey-based evaluation of simulated expert consent conversations by key stakeholders.
Zoe Atsaidis
Stephan Robitaille
Elena Guadagno
Jeffrey Wiseman
Sherif Emil
Dan Poenaru
2023-01-01
Journal of Pediatric Surgery (publié)
doi.org
BARVINN: Arbitrary Precision DNN Accelerator Controlled by a RISC-V CPU
Mohammadhossein Askarihemmat
Sean Wagner
O. Bilaniuk
Yassine Hariri
Yvon Savaria
J. David
We present a DNN accelerator that allows inference at arbitrary precision with dedicated processing elements that are configurable at the bi… (voir plus)t level. Our DNN accelerator has 8 Processing Elements controlled by a RISC-V controller with a combined 8.2 TMACs of computational power when implemented with the recent Alveo U250 FPGA platform. We develop a code generator tool that ingests CNN models in ONNX format and generates an executable com-mand stream for the RISC-V controller. We demonstrate the scalable throughput of our accelerator by running different DNN kernels and models when different quantization levels are selected. Compared to other low precision accelerators, our accelerator provides run time programmability without hardware reconfiguration and can accelerate DNNs with multiple quantization levels, regardless of the target FPGA size. BARVINN is an open source project and it is available at https://github.com/hossein1387/BARVINN.
2022-12-31
ArXiv (preprint)
doi.org
arxiv.org
Simplicity and learning to distinguish arguments from modifiers
Leon Bergen
E. Gibson
Timothy O'Donnell
2022-12-28
Journal of Language Modelling (publié)
doi.org
How programmers find online learning resources
Deeksha M. Arya
Jin Guo
Martin P. Robillard
2022-12-24
Empirical Software Engineering (publié)
doi.org
FaithDial: A Faithful Benchmark for Information-Seeking Dialogue
Nouha Dziri
Ehsan Kamalloo
Sivan Milton
Osmar Zaiane
Mo Yu
Edoardo Ponti
Siva Reddy
Abstract The goal of information-seeking dialogue is to respond to seeker queries with natural language utterances that are grounded on know… (voir plus)ledge sources. However, dialogue systems often produce unsupported utterances, a phenomenon known as hallucination. To mitigate this behavior, we adopt a data-centric solution and create FaithDial, a new benchmark for hallucination-free dialogues, by editing hallucinated responses in the Wizard of Wikipedia (WoW) benchmark. We observe that FaithDial is more faithful than WoW while also maintaining engaging conversations. We show that FaithDial can serve as training signal for: i) a hallucination critic, which discriminates whether an utterance is faithful or not, and boosts the performance by 12.8 F1 score on the BEGIN benchmark compared to existing datasets for dialogue coherence; ii) high-quality dialogue generation. We benchmark a series of state-of-the-art models and propose an auxiliary contrastive objective that achieves the highest level of faithfulness and abstractiveness based on several automated metrics. Further, we find that the benefits of FaithDial generalize to zero-shot transfer on other datasets, such as CMU-Dog and TopicalChat. Finally, human evaluation reveals that responses generated by models trained on FaithDial are perceived as more interpretable, cooperative, and engaging.
2022-12-23
Transactions of the Association for Computational Linguistics (publié)
doi.org
arxiv.org
Post-hoc Interpretability for Neural NLP: A Survey
Andreas Madsen
Siva Reddy
Sarath Chandar Anbil Parthipan
2022-12-23
ACM Computing Surveys (publié)
doi.org
arxiv.org
Towards Continual Reinforcement Learning: A Review and Perspectives
Khimya Khetarpal
Matthew D Riemer
Irina Rish
Doina Precup
2022-12-22
Journal of Artificial Intelligence Research (publié)
doi.org
arxiv.org
BLOOM+1: Adding Language Support to BLOOM for Zero-Shot Prompting
Zheng-Xin Yong
Hailey Schoelkopf
Niklas Muennighoff
Alham Fikri Aji
David Ifeoluwa Adelani
Khalid Almubarak
M. Saiful Bari
Lintang A. Sutawika
Jungo Kasai
Ahmed Baruwa
Genta Indra Winata
Stella Biderman
Dragomir R. Radev
Vassilina Nikoulina
The BLOOM model is a large publicly available multilingual language model, but its pretraining was limited to 46 languages. To extend the be… (voir plus)nefits of BLOOM to other languages without incurring prohibitively large costs, it is desirable to adapt BLOOM to new languages not seen during pretraining. In this work, we apply existing language adaptation strategies to BLOOM and benchmark its zero-shot prompting performance on eight new languages in a resource-constrained setting. We find language adaptation to be effective at improving zero-shot performance in new languages. Surprisingly, we find that adapter-based finetuning is more effective than continued pretraining for large models. In addition, we discover that prompting performance is not significantly affected by language specifics, such as the writing system. It is primarily determined by the size of the language adaptation data. We also add new languages to BLOOMZ, which is a multitask finetuned version of BLOOM capable of following task instructions zero-shot. We find including a new language in the multitask fine-tuning mixture to be the most effective method to teach BLOOMZ a new language. We conclude that with sufficient training data language adaptation can generalize well to diverse languages. Our code is available at https://github.com/bigscience-workshop/multilingual-modeling.
2022-12-19
ArXiv (prépublication)
doi.org
arxiv.org
Optimal Transport for Unsupervised Hallucination Detection in Neural Machine Translation
Nuno Miguel Guerreiro
Pierre Colombo
Pablo Piantanida
André Martins
Neural machine translation (NMT) has become the de-facto standard in real-world machine translation applications. However, NMT models can un… (voir plus)predictably produce severely pathological translations, known as hallucinations, that seriously undermine user trust. It becomes thus crucial to implement effective preventive strategies to guarantee their proper functioning. In this paper, we address the problem of hallucination detection in NMT by following a simple intuition: as hallucinations are detached from the source content, they exhibit encoder-decoder attention patterns that are statistically different from those of good quality translations. We frame this problem with an optimal transport formulation and propose a fully unsupervised, plug-in detector that can be used with any attention-based NMT model. Experimental results show that our detector not only outperforms all previous model-based detectors, but is also competitive with detectors that employ external models trained on millions of samples for related tasks such as quality estimation and cross-lingual sentence similarity.
2022-12-19
ArXiv (prépublication)
doi.org
arxiv.org
The KITMUS Test: Evaluating Knowledge Integration from Multiple Sources
Akshatha Arodi
Martin Pömsl
Kaheer Suleman
Adam Trischler
Alexandra Olteanu
Jackie Cheung
Many state-of-the-art natural language understanding (NLU) models are based on pretrained neural language models. These models often make in… (voir plus)ferences using information from multiple sources. An important class of such inferences are those that require both background knowledge, presumably contained in a model’s pretrained parameters, and instance-specific information that is supplied at inference time. However, the integration and reasoning abilities of NLU models in the presence of multiple knowledge sources have been largely understudied. In this work, we propose a test suite of coreference resolution subtasks that require reasoning over multiple facts. These subtasks differ in terms of which knowledge sources contain the relevant facts. We also introduce subtasks where knowledge is present only at inference time using fictional knowledge. We evaluate state-of-the-art coreference resolution models on our dataset. Our results indicate that several models struggle to reason on-the-fly over knowledge observed both at pretrain time and at inference time. However, with task-specific training, a subset of models demonstrates the ability to integrate certain knowledge types from multiple sources. Still, even the best performing models seem to have difficulties with reliably integrating knowledge presented only at inference time.
2022-12-15
ArXiv (prépublication)
doi.org
arxiv.org
Dynamic Consolidation for Continual Learning
Hang Li
Chen Ma
Xi Chen
Xue (Steve) Liu
Abstract Training deep learning models from a stream of nonstationary data is a critical problem to be solved to achieve general artificial … (voir plus)intelligence. As a promising solution, the continual learning (CL) technique aims to build intelligent systems that have the plasticity to learn from new information without forgetting the previously obtained knowledge. Unfortunately, existing CL methods face two nontrivial limitations. First, when updating a model with new data, existing CL methods usually constrain the model parameters within the vicinity of the parameters optimized for old data, limiting the exploration ability of the model; second, the important strength of each parameter (used to consolidate the previously learned knowledge) is fixed and thus is suboptimal for the dynamic parameter updates. To address these limitations, we first relax the vicinity constraints with a global definition of the important strength, which allows us to explore the full parameter space. Specifically, we define the important strength as the sensitivity of the global loss function to the model parameters. Moreover, we propose adjusting the important strength adaptively to align it with the dynamic parameter updates. Through extensive experiments on popular data sets, we demonstrate that our proposed method outperforms the strong baselines by up to 24% in terms of average accuracy.
2022-12-14
Neural Computation (publié)
doi.org
A Tweedie Compound Poisson Model in Reproducing Kernel Hilbert Space
Yi Lian
Archer Yang
Boxiang Wang
Peng Shi
Robert William Platt
Abstract Tweedie models can be used to analyze nonnegative continuous data with a probability mass at zero. There have been wide application… (voir plus)s in natural science, healthcare research, actuarial science, and other fields. The performance of existing Tweedie models can be limited on today’s complex data problems with challenging characteristics such as nonlinear effects, high-order interactions, high-dimensionality and sparsity. In this article, we propose a kernel Tweedie model, Ktweedie, and its sparse variant, SKtweedie, that can simultaneously address the above challenges. Specifically, nonlinear effects and high-order interactions can be flexibly represented through a wide range of kernel functions, which is fully learned from the data; In addition, while the Ktweedie can handle high-dimensional data, the SKtweedie with integrated variable selection can further improve the interpretability. We perform extensive simulation studies to justify the prediction and variable selection accuracy of our method, and demonstrate the applications in ratemaking and loss-reserving in general insurance. Overall, the Ktweedie and SKtweedie outperform existing Tweedie models when there exist nonlinear effects and high-order interactions, particularly when the dimensionality is high relative to the sample size. The model is implemented in an efficient and user-friendly R package ktweedie (https://cran.r-project.org/package=ktweedie).
2022-12-13
Technometrics (publié)
doi.org