Publications

Tequila: Deadzone-free Ternary Quantization for Large Language Models
Hong Huang
Decheng Wu
Rui Cen
Guanghua Yu
Zonghang Li
Kai Liu
Jianchen Zhu
Peng Chen
Xue Liu
Dapeng Wu
Quantization techniques are essential for the deployment of Large Language Models (LLMs) on edge devices. However, prevailing methods often … (voir plus)rely on mixed-precision multiplication that lacks efficient hardware support, making it not feasible. Ternary weight quantization addresses this by constraining weights to {-1, 0, 1}, replacing expensive multiplications with hardware-efficient additions. However, such aggressive compression leads to significant accuracy degradation, even after costly quantization-aware training with massive data. We identify the core issue as _**deadzone trapping**: a large number of weights are trapped at the deadzone boundary._ This occurs because these weights receive only noisy, less informative gradients, preventing stable escape from the deadzone and severely impeding model capacity and optimization. To address this issue, we propose **Tequila**, a trapping-free quantization optimization method that reactivates deadzone-trapped weights by repurposing them as dynamic biases. This allows the repurposed weights to provide a continuous signal in the forward pass and, critically, receive direct, meaningful gradient signals during backpropagation, thereby enhancing model capacity and optimization with nearly _zero_ inference overhead. Extensive evaluations demonstrate that Tequila outperforms state-of-the-art (SOTA) ternary quantization methods across five benchmarks. Specifically, on the ARC benchmark, it achieves
2025-12-31
International Conference on Learning Representations (Accept (Poster))
openreview.net
openreview.net
The 2nd Workshop on Foundation Models for Science: Real-World Impact and Science-First Design
Wuyang Chen
Yongji Wang
N. Benjamin Erichson
Laurence Perreault-Levasseur
Bo Li
Damian Borth
Swarat Chaudhuri
Scientific foundation models should be built for science, not for generic AI tastes or leaderboard prestige. This workshop centers problem-d… (voir plus)riven design: models that measurably advance real scientific inquiries, e.g., forecasting extreme climate events, accelerating materials discovery, understanding biological mechanisms, co-developed with domain experts and validated against field data, experiments, and downstream impact. We argue that foundation models for science must be built differently from language and vision. Scientific data are physical, causal, spatiotemporal, and often scarce or biased; objectives must reflect mechanistic fidelity, not just predictive accuracy. This calls for scientific priors and constraints, robust uncertainty quantification (UQ), and architectures that natively handle multi-modality (e.g., grids, meshes, spectra, time series, point clouds, text, images, code). It also demands tight integration with classical scientific tools (simulators, PDE solvers, optimization and inference engines, and HPC workflows) to yield hybrid systems that are faster, more accurate, and more trustworthy. We will highlight opportunities and hard problems unique to science: enforcing conservation laws and symmetries; learning across vast spatial and temporal scales; representing extreme events and tipping points; calibrating and validating UQ; and developing evaluation protocols that reward mechanistic insight and actionable reliability. The goal is a roadmap for building, training, and deploying scientific foundation models that accelerate discovery while respecting the structure of the natural world.
2025-12-31
Workshop Proposals @ International Conference on Learning Representations (publié)
openreview.net
openreview.net
The Design Space of Tri-Modal Masked Diffusion Models
Louis Bethune
Victor Turrisi
Bruno Kacper Mlodozeniec
Pau Rodriguez Lopez
Lokesh Boominathan
Nikhil Bhendawade
Amitis Shidani
Joris Pelemans
Theo X. Olausson
Devon Hjelm
Paul Dixon
Joao Monteiro
Pierre Ablin
Vishnu Banna
Arno Blaas
Nick Henderson
Kari Noriy
Dan Busbridge
Josh Susskind
Marco Cuturi … (voir 4 de plus)
Irina Belousova
Luca Zappella
Russ Webb
Jason Ramapuram
Discrete diffusion models have emerged as strong alternatives to autoregressive language models, with recent work initializing and fine-tuni… (voir plus)ng a base unimodal model for bimodal generation. Diverging from previous approaches, we introduce the first tri-modal masked diffusion model pretrained from scratch on text, image-text, and audio-text data. We systematically analyze multimodal scaling laws, modality mixing ratios, noise schedules, and batch-size effects, and we provide optimized inference sampling defaults. Our batch-size analysis yields a novel stochastic differential equation (SDE)-based reparameterization that eliminates the need for tuning the optimal batch size as reported in recent work. This reparameterization decouples the physical batch size, often chosen based on compute constraints (GPU saturation, FLOP efficiency, wall-clock time), from the logical batch size, chosen to balance gradient variance during stochastic optimization. Finally, we pretrain a preliminary 3B-parameter tri-modal model on 6.4T tokens, demonstrating the capabilities of a unified design and achieving strong results in text generation, text-to-image tasks, and text-to-speech tasks. Our work represents the largest-scale systematic open study of multimodal discrete diffusion models conducted to date, providing insights into scaling behaviors across multiple modalities.
2025-12-31
arXiv (prépublication)
doi.org
arxiv.org
At the Edge of Understanding: Sparse Autoencoders Trace The Limits of Transformer Generalization
Praneet Suresh
Jack Stanley
Sonia Joseph
Luca Scimeca
Danilo Bzdok
Pre-trained transformers have demonstrated remarkable generalization abilities, at times extending beyond the scope of their training data. … (voir plus)Yet, real-world deployments often face unexpected or adversarial data that diverges from training data distributions. Without explicit mechanisms for handling such shifts, model reliability and safety degrade, urging more disciplined study of out-of-distribution (OOD) settings for transformers. By systematic experiments, we present a mechanistic framework for delineating the precise contours of transformer model robustness. We find that OOD inputs, including subtle typos and jailbreak prompts, drive language models to operate on an increased number of fallacious concepts in their internals. We leverage this device to quantify and understand the degree of distributional shift in prompts, enabling a mechanistically grounded fine-tuning strategy to robustify LLMs. Expanding the very notion of OOD from input data to a model’s private computational processes—a new transformer diagnostic at inference time—is a critical step toward making AI systems safe for deployment across science, business, and government.
2025-12-31
International Conference on Machine Learning (Accept (regular))
openreview.net
openreview.net
The Expressive Limits of Diagonal SSMs for State-Tracking
Mehran Shakerinava
Behnoush Khavari
Siamak Ravanbakhsh
Sarath Chandar
State-Space Models (SSMs) have recently been shown to achieve strong empirical performance on a variety of long-range sequence modeling task… (voir plus)s while remaining efficient and highly-parallelizable. However, the theoretical understanding of their expressive power remains limited. In this work, we study the expressivity of input-Dependent Complex-valued Diagonal (DCD) State-Space Models (SSMs) on sequential state-tracking tasks for abstract groups. It is easy to show that a single DCD SSM layer with a universal decoder can track any Abelian group at finite precision by decomposing it into a product of cyclic groups. We show that this is tight by proving that such a model cannot track any non-Abelian group at finite precision. We further establish the expressivity of multi-layer DCD SSMs. We show that a
2025-12-31
International Conference on Learning Representations (Accept (Poster))
doi.org
openreview.net
The Future of the AI Summit Series: One Year Later
Lucia Velasco
Yoshua Bengio
Robert Trager
Joslyn Barnhart
Julia Morse
Roxana Radu
Ranjit Lall
Stephen Casper
Charles Martinet
2025-12-31
SSRN Electronic Journal (accepté)
doi.org
The Geometry and Topology of Circuits: the Manifolds of Modular Addition
Gabriela Moisescu-Pareja
Gavin McCracken
Harley Wiltzer
Vincent Létourneau
Colin Daniels
Doina Precup
Jonathan Love
The Clock and Pizza interpretations, associated with architectures differing in either uniform or learnable attention, were introduced to ar… (voir plus)gue that different architectural designs can yield distinct circuits for modular addition. In this work, we show that this is not the case, and that both the uniform and trainable attention architectures implement the same algorithm via topologically and geometrically equivalent representations. Our methodology goes beyond the interpretation of individual neurons and weights. Instead, we identify all of the neurons corresponding to each learned representation and then study the collective group of neurons as one entity. This method reveals that each learned representation is a manifold that we can study utilizing tools from topology. Based on this insight, we can statistically analyze the learned representations across hundreds of circuits to demonstrate the similarity between learned modular addition circuits that arise naturally from common deep learning paradigms.
2025-12-31
International Conference on Learning Representations (Accept (Poster))
openreview.net
openreview.net
The Impact of Pediatric Surgery Global Travel Fellowships: A Study by the Canadian Association of Paediatric Surgeons Global Partnership Committee.
Sacha Williams
Natasha Bejjani
Elena Guadagno
Robert Baird
Shahrzad Joharifard
Melanie Morris
Robin Petroze
Dan Poenaru
Sherif Emil
2025-12-31
Journal of Pediatric Surgery (publié)
doi.org
The Intricate Dance of Prompt Complexity, Quality, Diversity and Consistency in T2I Models
Zhang Xiaofeng
Aaron Courville
Michal Drozdzal
Adriana Romero-Soriano
Text-to-image (T2I) models offer great potential for creating virtually limitless synthetic data, a valuable resource compared to fixed and … (voir plus)finite real datasets. Previous works evaluate the utility of synthetic data from T2I models on three key desiderata: quality, diversity, and consistency. While prompt engineering is the primary means of interacting with T2I models, the systematic impact of prompt complexity on these critical utility axes remains underexplored. In this paper, we first conduct synthetic experiments to motivate the difficulty of generalization w.r.t. prompt complexity and explain the observed difficulty with theoretical derivations. Then, we introduce a new evaluation framework that can compare the utility of real data and synthetic data, and present a comprehensive analysis of how prompt complexity influences the utility of synthetic data generated by commonly used T2I models. We conduct our study across diverse datasets, including CC12M, ImageNet-1k, and DCI, and evaluate different inference-time intervention methods. Our synthetic experiments show that generalizing to more general conditions is harder than the other way round, since the former needs an estimated likelihood that is not learned by diffusion models. Our large-scale empirical experiments reveal that increasing prompt complexity results in lower conditional diversity and prompt consistency, while reducing the synthetic-to-real distribution shift, which aligns with the synthetic experiments. Moreover, current inference-time interventions can augment the diversity of the generations at the expense of moving outside the support of real data. Among those interventions, prompt expansion, by deliberately using a pre-trained language model as a likelihood estimator, consistently achieves the highest performance in both image diversity and aesthetics, even higher than that of real data. Combining advanced guidance interventions with prompt expansion results in the most appealing utility trade-offs of synthetic data.
2025-12-31
International Conference on Learning Representations (Accept (Poster))
openreview.net
openreview.net
The Markovian Thinker
Milad Aghajohari
Kamran Chitsaz
Amirhossein Kazemnejad
Sarath Chandar
Alessandro Sordoni
Aaron Courville
Siva Reddy
Reasoning LLMs suffer from quadratic compute growth as their context length increases, making reinforcement learning with verifiable rewards… (voir plus) (RLVR) and test-time scaling prohibitively expensive. Prior work has tried to lighten the computational burden by shortening reasoning traces through pruning, summarization, or multi-stage training, but these methods remain bound to quadratic costs. We introduce Delethink, a thinking algorithm that realizes the Markovian Thinking Paradigm. Instead of producing one long monolithic reasoning trace, Delethink thinks in a sequence of chunks, the Delethink trace. Each chunk continues reasoning by referring only to a fixed number of prior tokens, which functions as a Markovian state sufficient for progressing reasoning, while deleting the rest. This preserves continuity without carrying the quadratic baggage. As a result, compute scales linearly and peak memory remains constant. In experiments, we show that Delethink can be applied directly to off-the-shelf reasoning models ranging from
2025-12-31
International Conference on Learning Representations (Accept (Poster))
openreview.net
openreview.net
On the sensitivity of SAR C- and L-band dual-polarized data for detection of early deforestation in the tropics
Africa I. Flores-Anderson
Jeffrey A. Cardille
Josef Kellndorfer
Franz J. Meyer
Pontus Olofsson
2025-12-31
Remote Sensing of Environment (publié)
doi.org
The Silent Thought: Modeling Internal Cognition in Full-Duplex Spoken Dialogue Models via Latent Reasoning
Donghang Wu
Tianyu Zhang
Yuxin Li
Hexin Liu
Chen Chen
Eng Siong Chng
Yoshua Bengio
During conversational interactions, humans subconsciously engage in concurrent thinking while listening to a speaker. Although this internal… (voir plus) cognitive processing may not always manifest as explicit linguistic structures, it is instrumental in formulating high-quality responses. Inspired by this cognitive phenomenon, we propose a novel **F**ull-duplex **LA**tent and **I**nternal **R**easoning method named FLAIR that conducts *latent* thinking simultaneously with speech perception. Unlike conventional "thinking" mechanisms in NLP, which require post-hoc generation, our approach aligns seamlessly with spoken dialogue systems: during the user’s speaking phase, it recursively feeds the latent embedding output from the previous step into the next step, enabling continuous reasoning that strictly adheres to causality without introducing additional latency. To enable this latent reasoning, we design an Evidence Lower Bound-based objective that supports efficient supervised finetuning via teacher forcing, circumventing the need for explicit reasoning annotations. Experiments demonstrate the effectiveness of this think-while-listening design, which achieves competitive results on a range of speech benchmarks. Furthermore, FLAIR robustly handles conversational dynamics and attains competitive performance on full-duplex interaction metrics.
2025-12-31
International Conference on Machine Learning (Accept (regular))
doi.org
openreview.net