Publications

The Effect of Data Corruption on Multimodal Long Form Responses
Daniel Z Kaplan
Alexis Roger
Mohamed Osman
Irina Rish
Despite significant progress, Vision-Language Models (VLMs) still struggle with hallucinations, especially in long-form responses. Existing … (voir plus)strategies have had limited successes in specific cases, and long-form generation remains problematic. In this work we attempt to establish the link between the data used to train the model and the hallucinations in the model's output. To this end, we examine hallucinations through data corruption. We develop a method to corrupt training data and then train models with this data to see the effect on performance. We will show that corrupting only a small portion of the long-form training data significantly impairs the performance of the model on long-form tasks, while leaving simpler tasks like visual question-answering and multiple choice relatively intact. All training code and models are released for reproducibility and future research.
2024-07-02
ICML.cc/2024/Workshop/FM-Wild (poster)
openreview.net
openreview.net
TriLM vs FloatLM: Ternary LLMs are more Performant than Quantized FP16 LLMs
Ayush Kaushal
Tejas Vaidhya
Tejas Pandey
Aaryan Bhagat
Irina Rish
Ternary LLMs offer significantly better performance for their size (measured in bits) than the models trained and deployed in FP16/BF16. Giv… (voir plus)en the widespread usage of quantization before deployment and advancements in Post Training Quantization of LLMs, a pivotal question arises: do ternary LLMs indeed provide any discernible benefits? To address this, we first build an open family of pre-trained ternary Large Language Models (TriLM). Additionally, we include their counterparts pre-trained in FP16 (FloatLM) and quantized versions of FloatLM (QuantLM) with parameters across almost two orders of magnitude - from 99M to 3.9B parameters. We demonstrate that TriLMs with 3B+ parameters start to offer competitive performance compared to FloatLMs with the same parameter count, while providing significantly better performance for their size. Specifically, TriLM 3.9B, with less bits than FloatLM 830M, ranks between FloatLM 2.4B and FloatLM 3.9B when averaged across 6 popular commonsense and reasoning benchmarks. TriLMs also outperform quantized models, with TriLM 3.9B surpassing the larger QuantLM-3bit 3.9B. Furthermore, across knowledge-based benchmarks, TriLM maintains a superiority for its size, but lags for its parameter count. TriLM 3.9B falls halfway between FloatLM 1.5B and 2.4B, close to QuantLM-4bit 2.4B. To advance research on Ternary LMs, we open source over 500+ checkpoints across the model families.
2024-07-02
ICML.cc/2024/Workshop/FM-Wild (poster)
openreview.net
openreview.net
VFA: Vision Frequency Analysis of Foundation Models and Human
Mohammad Javad Darvishi Bayazi
Md Rifat Arefin
Jocelyn Faubert
Irina Rish
2024-07-02
ICML.cc/2024/Workshop/FM-Wild (poster)
doi.org
openreview.net
A Bayesian Non-Stationary Heteroskedastic Time Series Model for Multivariate Critical Care Data
Zayd Omar
David A. Stephens
Alexandra M. Schmidt
David L Buckeridge
2024-07-01
Statistics in Medicine (publié)
doi.org
arxiv.org
Temperature-dependent Spike-ACE2 interaction of Omicron subvariants is associated with viral transmission
Mehdi Benlarbi
Shilei Ding
Étienne Bélanger
Alexandra Tauzin
Raphaël Poujol
Halima Medjahed
Omar El Ferri
Yuxia Bo
Catherine Bourassa
Julie Hussin
Judith Fafard
Marzena Pazgier
Inès Levade
Cameron Abrams
Marceline Côté
Andrés Finzi
The continued evolution of severe acute respiratory syndrome 2 (SARS-CoV-2) requires persistent monitoring of its subvariants. Omicron subva… (voir plus)riants are responsible for the vast majority of SARS-CoV-2 infections worldwide, with XBB and BA.2.86 sublineages representing more than 90% of circulating strains as of January 2024. To better understand parameters involved in viral transmission, we characterized the functional properties of Spike glycoproteins from BA.2.75, CH.1.1, DV.7.1, BA.4/5, BQ.1.1, XBB, XBB.1, XBB.1.16, XBB.1.5, FD.1.1, EG.5.1, HK.3, BA.2.86 and JN.1. We tested their capacity to evade plasma-mediated recognition and neutralization, binding to angiotensin-converting enzyme 2 (ACE2), their susceptibility to cold inactivation, Spike processing, as well as the impact of temperature on Spike-ACE2 interaction. We found that compared to the early wild-type (D614G) strain, most Omicron subvariants' Spike glycoproteins evolved to escape recognition and neutralization by plasma from individuals who received a fifth dose of bivalent (BA.1 or BA.4/5) mRNA vaccine and improve ACE2 binding, particularly at low temperatures. Moreover, BA.2.86 had the best affinity for ACE2 at all temperatures tested. We found that Omicron subvariants’ Spike processing is associated with their susceptibility to cold inactivation. Intriguingly, we found that Spike-ACE2 binding at low temperature was significantly associated with growth rates of Omicron subvariants in humans. Overall, we report that Spikes from newly emerged Omicron subvariants are relatively more stable and resistant to plasma-mediated neutralization, present improved affinity for ACE2 which is associated, particularly at low temperatures, with their growth rates. The persistent evolution of SARS-CoV-2 gave rise to a wide range of variants harboring new mutations in their Spike glycoproteins. Several factors have been associated with viral transmission and fitness such as plasma-neutralization escape and ACE2 interaction. To better understand whether additional factors could be of importance in SARS-CoV-2 variants’ transmission, we characterize the functional properties of Spike glycoproteins from several Omicron subvariants. We found that the Spike glycoprotein of Omicron subvariants presents an improved escape from plasma-mediated recognition and neutralization, Spike processing, and ACE2 binding which was further improved at low temperature. Intriguingly, Spike-ACE2 interaction at low temperature is strongly associated with viral growth rate, as such, low temperatures could represent another parameter affecting viral transmission.
2024-07-01
mBio (publié)
doi.org
089 Levers and limitations of artificial intelligence (AI) to support the assessment and implementation of shared decision making (SDM): perspectives of key stakeholders
Anik Giguère
Adrian Edwards
Denitza Williams
France Légaré
Natalie Joseph-Williams
Samira Abbasgholizadeh-Rahimi
Karina Prévost
Marie-Clare Hunter
Anna Torrens-Burton
Justine Laloux
2024-06-30
International Symposium on Data Mining (publié)
doi.org
169Yb-based high dose rate intensity modulated brachytherapy for focal treatment of prostate cancer
Maude Robitaille
Cynthia Ménard
Gabriel Famulari
Dominic Béliveau-Nadeau
S. Enger
2024-06-30
Brachytherapy (publié)
doi.org
Accelerated Benders Decomposition and Local Branching for Dynamic Maximum Covering Location Problems
Steven Lamontagne
Margarida Carvalho
Ribal Atallah
The maximum covering location problem (MCLP) is a key problem in facility location, with many applications and variants. One such variant is… (voir plus) the dynamic (or multi-period) MCLP, which considers the installation of facilities across multiple time periods. To the best of our knowledge, no exact solution method has been proposed to tackle large-scale instances of this problem. To that end, in this work, we expand upon the current state-of-the-art branch-and-Benders-cut solution method in the static case, by exploring several acceleration techniques. Additionally, we propose a specialised local branching scheme, that uses a novel distance metric in its definition of subproblems and features a new method for efficient and exact solving of the subproblems. These methods are then compared through extensive computational experiments, highlighting the strengths of the proposed methodologies.
2024-06-30
Computers & Operations Research (publié)
doi.org
arxiv.org
Deep learning based vessel arrivals monitoring via autoregressive statistical control charts
Sara El Mekkaoui
Ghait Boukachab
Loubna Benabbou
Abdelaziz Berrado
2024-06-30
WMU Journal of Maritime Affairs (JoMA) (publié)
doi.org
Imagining a Future of Designing with AI: Dynamic Grounding, Constructive Negotiation, and Sustainable Motivation
Priyan Vaithilingam
Ian Arawjo
Elena L. Glassman
2024-06-30
Designing Interactive Systems Conference (publié)
doi.org
arxiv.org
Do LLMs Meet the Needs of Software Tutorial Writers? Opportunities and Design Implications
Avinash Bhat
Disha Shrivastava
Jin L.C. Guo
Creating software tutorials involves developing accurate code examples and explanatory text that engages and informs the reader. Large Langu… (voir plus)age Models (LLMs) demonstrate a strong capacity to generate both text and code, but their potential to assist tutorial writing is unknown. By interviewing and observing seven experienced writers using OpenAI playground as an exploration environment, we uncover design opportunities for leveraging LLMs in software tutorial writing. Our findings reveal background research, resource creation, and maintaining quality standards as critical areas where LLMs could significantly assist writers. We observe how tutorial writers generated tutorial content while exploring LLMs’ capabilities, formulating prompts, verifying LLM outputs, and reflecting on interaction goals and strategies. Our observation highlights that the unpredictability of LLM outputs and unintuitive interface design contributed to skepticism about LLM’s utility. Informed by these results, we contribute recommendations for designing LLM-based tutorial writing tools to mitigate usability challenges and harness LLMs’ full potential.
2024-06-30
Conference on Designing Interactive Systems (publié)
doi.org
A logistics provider’s profit maximization facility location problem with random utility maximizing followers
David Pinzon Ulloa
Emma Frejinger
Bernard Gendron
2024-06-30
Computers & Operations Research (publié)
doi.org
arxiv.org