Publications

Who Guards the Guardians? The Challenges of Evaluating Identifiability of Learned Representations
Shruti Joshi
Theo Saulus
Wieland Brendel
Philippe Brouillard
Dhanya Sridhar
Patrik Reizinger
Identifiability in representation learning is commonly evaluated using standard metrics (e.g., MCC, DCI, R^2) on synthetic benchmarks with k… (voir plus)nown ground-truth factors. These metrics are assumed to reflect recovery up to the equivalence class guaranteed by identifiability theory. We show that this assumption holds only under specific structural conditions: each metric implicitly encodes assumptions about both the data-generating process (DGP) and the encoder. When these assumptions are violated, metrics become misspecified and can produce systematic false positives and false negatives. Such failures occur both within classical identifiability regimes and in post-hoc settings where identifiability is most needed. We introduce a taxonomy separating DGP assumptions from encoder geometry, use it to characterise the validity domains of existing metrics, and release an evaluation suite for reproducible stress testing and comparison.
2026-02-26
arXiv (prépublication)
doi.org
arxiv.org
CellPace: A temporal diffusion-forcing framework for simulation, interpolation and forecasting of single-cell dynamics
Chen Su
Amin Emad
Abstract Single-cell omics technologies resolve cellular heterogeneity at high resolution but provide only static snapshots of continuous de… (voir plus)velopmental processes. This makes it difficult to recover coherent temporal dynamics when developmental stages are irregularly sampled or missing. While recent generative models can simulate observed cell states, they often treat timepoints as discrete categories, hindering interpolation across gaps and extrapolation to unobserved future stages. We present CellPace, a generative model that learns and generates developmental dynamics by leveraging a transformer-based temporal diffusion backbone conditioned on continuous, gap-aware temporal encodings. Across diverse mouse developmental lineages, CellPace achieves state-of-the-art performance in simulation, interpolation, and forecasting tasks. Beyond recovering global population statistics, generated cells preserve fine-grained biological structure, retaining dynamic gene regulatory programs and mapping accurately to anatomical regions in spatial transcriptomics data. Furthermore, CellPace extends naturally to multi-modal data, modeling joint RNA-chromatin dynamics even when temporal ordering is inferred from pseudotime. Together, these results position CellPace as a robust framework for modeling and generating continuous developmental dynamics from sparse, cross-sectional single-cell data.
2026-02-25
bioRxiv (prépublication)
doi.org
InnerQ: Hardware-aware Tuning-free Quantization of KV Cache for Large Language Models
Mohammadreza Tayaranian
Amir Ardakani
Warren J. Gross
Reducing the hardware footprint of large language models (LLMs) during decoding is critical for efficient long-sequence generation. A key bo… (voir plus)ttleneck is the key-value (KV) cache, whose size scales with sequence length and easily dominates the memory footprint of the model. Previous work proposed quantization methods that are focused on compressing the KV cache while maintaining its information. We introduce InnerQ, a hardware-aware KV-cache quantization scheme that lowers decode latency without sacrificing accuracy. InnerQ applies group-wise quantization while grouping the cache matrices over their inner dimension. Unlike previous work that group over the outer dimension, InnerQ aligns dequantization with the vector-matrix multiplication and enables scale factor reuse across GPU compute units. This reduces memory accesses and accelerates dequantization, yielding up to
2026-02-25
arXiv (prépublication)
doi.org
arxiv.org
Soil microbiome prediction using traditional machine learning and deep learning models
Zahia Aouabed
Vincent Therrien
Mohamed Achraf Bouaoune
Mohammadreza Bakhtyari
Mohamed Hijri
Vladimir Makarenkov
The accuracy of macrobiological community predictions largely depends on the taxonomic scale considered. Nowadays, the applicability of such… (voir plus) predictions remains an important challenge when extended to microbial soil communities. This is not only due to the lack of reliable benchmark data, but also to a greater diversity of the soil microorganisms compared to other environments. In this study, we use six traditional machine learning regression models and one deep learning regressor to predict relative frequencies of bacterial and fungal communities within the soil microbiome based on environmental factors. We analyze the data from two publicly available soil microbiome datasets: (1) Data collected by Averill and co-authors and analyzed in a recent Nature Ecology and Evolution article, and (2) Data extracted from the NEON database, to estimate the composition of bacterial and fungal communities at the functional (i.e. functional group level) and taxonomic scales (i.e. phylum, class, order, family, and genus levels). Our findings suggest the presence of a general pattern across the observed taxonomic scales according to which the predictability of the soil microbiome increases with taxonomic scale. However, a notable exception occurs when machine learning models are applied to predict bacterial communities at the functional group level for Averill et al.’s data when all of them fail to provide accurate predictions results. The best overall results obtained include the value of the coefficient of determination
2026-02-24
Scientific Reports (publié)
doi.org
Soil microbiome prediction using traditional machine learning and deep learning models
Zahia Aouabed
Vincent Therrien
Mohamed Achraf Bouaoune
Mohammadreza Bakhtyari
Mohamed Hijri
Vladimir Makarenkov
The accuracy of macrobiological community predictions largely depends on the taxonomic scale considered. Nowadays, the applicability of such… (voir plus) predictions remains an important challenge when extended to microbial soil communities. This is not only due to the lack of reliable benchmark data, but also to a greater diversity of the soil microorganisms compared to other environments. In this study, we use six traditional machine learning regression models and one deep learning regressor to predict relative frequencies of bacterial and fungal communities within the soil microbiome based on environmental factors. We analyze the data from two publicly available soil microbiome datasets: (1) Data collected by Averill and co-authors and analyzed in a recent Nature Ecology and Evolution article, and (2) Data extracted from the NEON database, to estimate the composition of bacterial and fungal communities at the functional (i.e. functional group level) and taxonomic scales (i.e. phylum, class, order, family, and genus levels). Our findings suggest the presence of a general pattern across the observed taxonomic scales according to which the predictability of the soil microbiome increases with taxonomic scale. However, a notable exception occurs when machine learning models are applied to predict bacterial communities at the functional group level for Averill et al.’s data when all of them fail to provide accurate predictions results. The best overall results obtained include the value of the coefficient of determination
2026-02-24
Scientific Reports (publié)
doi.org
Detoxifying LLMs via Representation Erasure-Based Preference Optimization
Nazanin Mohammadi Sepahvand
Eleni Triantafillou
Hugo Larochelle
Doina Precup
Daniel M. Roy
Gintare Karolina Dziugaite
Large language models (LLMs) trained on webscale data can produce toxic outputs, raising concerns for safe deployment. Prior defenses, based… (voir plus) on applications of DPO, NPO, and similar algorithms, reduce the likelihood of harmful continuations, but not robustly so: they are vulnerable to adversarial prompting and easily undone by fine-tuning-based relearning attacks. Indeed, research has shown that these edits to the model are superficial: linear probing reveals that harmful "directions" remain present in representations. To address this, we propose Representation Erasure-based Preference Optimization (REPO), reformulating detoxification as a token-level preference problem. Using a novel objective with preference data, we force the representations of toxic continuations to converge toward their benign counterparts. Our mechanistic analysis reveals that this granular approach is critical: unlike baselines, REPO induces deep, localized edits to toxicity-encoding neurons while preserving general model utility. Exhaustive evaluations show that REPO achieves state-of-the-art robustness, stopping sophisticated threats-including relearning attacks and enhanced GCG jailbreaks-where existing representation- and output-based methods fail.
2026-02-23
arXiv (prépublication)
doi.org
arxiv.org
WildSVG: Towards Reliable SVG Generation Under Real-Word Conditions
Marco Terral
Haotian Zhang
Tianyang Zhang
Meng Lin
Xiaoqing Xie
Haoran Dai
Darsh Kaushik
Pai Peng
Nicklas Scharpff
David Vázquez
Joan Rodriguez
We introduce the task of SVG extraction, which consists in translating specific visual inputs from an image into scalable vector graphics. E… (voir plus)xisting multimodal models achieve strong results when generating SVGs from clean renderings or textual descriptions, but they fall short in real-world scenarios where natural images introduce noise, clutter, and domain shifts. A central challenge in this direction is the lack of suitable benchmarks. To address this need, we introduce the WildSVG Benchmark, formed by two complementary datasets: Natural WildSVG, built from real images containing company logos paired with their SVG annotations, and Synthetic WildSVG, which blends complex SVG renderings into real scenes to simulate difficult conditions. Together, these resources provide the first foundation for systematic benchmarking SVG extraction. We benchmark state-of-the-art multimodal models and find that current approaches perform well below what is needed for reliable SVG extraction in real scenarios. Nonetheless, iterative refinement methods point to a promising path forward, and model capabilities are steadily improving
2026-02-23
arXiv (prépublication)
doi.org
arxiv.org
Grokking Finite-Dimensional Algebra
Pascal Jr Tikeng Notsawo
Guillaume Dumas
Guillaume Rabusseau
This paper investigates the grokking phenomenon, which refers to the sudden transition from a long memorization to generalization observed d… (voir plus)uring neural networks training, in the context of learning multiplication in finite-dimensional algebras (FDA). While prior work on grokking has focused mainly on group operations, we extend the analysis to more general algebraic structures, including non-associative, non-commutative, and non-unital algebras. We show that learning group operations is a special case of learning FDA, and that learning multiplication in FDA amounts to learning a bilinear product specified by the algebra's structure tensor. For algebras over the reals, we connect the learning problem to matrix factorization with an implicit low-rank bias, and for algebras over finite fields, we show that grokking emerges naturally as models must learn discrete representations of algebraic elements. This leads us to experimentally investigate the following core questions: (i) how do algebraic properties such as commutativity, associativity, and unitality influence both the emergence and timing of grokking, (ii) how structural properties of the structure tensor of the FDA, such as sparsity and rank, influence generalization, and (iii) to what extent generalization correlates with the model learning latent embeddings aligned with the algebra's representation. Our work provides a unified framework for grokking across algebraic structures and new insights into how mathematical structure governs neural network generalization dynamics.
2026-02-22
arXiv (prépublication)
doi.org
arxiv.org
Learning to Solve Complex Problems via Dataset Decomposition
WANRU ZHAO
Lucas Caccia
Zhengyan Shi
Minseon Kim
Weijia Xu
Alessandro Sordoni
Curriculum learning is a class of training strategies that organizes the data being exposed to a model by difficulty, gradually from simpler… (voir plus) to more complex examples. This research explores a reverse curriculum generation approach that recursively decomposes complex datasets into simpler, more learnable components. We propose a teacher-student framework where the teacher is equipped with the ability to reason step-by-step, which is used to recursively generate easier versions of examples, enabling the student model to progressively master difficult tasks. We propose a novel scoring system to measure data difficulty based on its structural complexity and conceptual depth, allowing curriculum construction over decomposed data. Experiments on math datasets (MATH and AIME) and code generation datasets demonstrate that models trained with curricula generated by our approach exhibit superior performance compared to standard training on original datasets.
2026-02-22
arXiv (prépublication)
doi.org
arxiv.org
Celo2: Towards Learned Optimization Free Lunch
Abhinav Moudgil
Boris Knyazev
Eugene Belilovsky
Learned optimizers are powerful alternatives to hand-designed update rules like Adam, yet they have seen limited practical adoption since th… (voir plus)ey often fail to meta-generalize beyond their training distribution and incur high meta-training cost. For instance, prior work, VeLO, scaled meta-training to 4,000 TPU months (
2026-02-21
arXiv (prépublication)
doi.org
arxiv.org
Panorama of Soft Tissue Tumours at a Tertiary Care Centre in Bihar: A Retrospective Observational Study
Ankit Anand
Vibhuti Kumar
Objective and Aim: Soft tissue tumors (STTs) represent a heterogeneous group of neoplasms with diverse histogenesis, biological behavior, an… (voir plus)d clinical outcomes. The present study aims to evaluate the spectrum, frequency, demographic distribution, anatomical location, and histopathological patterns of soft tissue tumors diagnosed at a tertiary care center in Bihar, India, with special emphasis on benign–malignant correlation and clinicopathological characteristics. Materials and Methods: This retrospective observational study was conducted in the Department of Pathology at a tertiary care teaching hospital in Bihar over a period of five years (January 2019–December 2023). All histopathologically confirmed cases of soft tissue tumors were included. Tumors were classified according to the WHO Classification of Soft Tissue and Bone Tumors (2020). Statistical analysis was performed using SPSS version 26.0. Descriptive statistics, chi-square test, and logistic regression analysis were applied. Results: A total of 312 cases of soft tissue tumors were analyzed. Benign tumors constituted 76.9%, intermediate tumors 7.4%, and malignant tumors 15.7%. The most common benign tumor was lipoma (38.1%), while undifferentiated pleomorphic sarcoma (21.4%) was the most frequent malignant tumor. Malignant tumors were significantly associated with age >40 years (p 0.001) and deep-seated location (p = 0.002). Conclusion: Soft tissue tumors in Bihar show a predominance of benign lesions with lipoma bei
2026-02-21
International Journal of Current Pharmaceutical Review and Research (publié)
doi.org
Stable Deep Reinforcement Learning via Isotropic Gaussian Representations
Ali Saheb Pasand
Johan Obando-Ceron
Aaron Courville
Pouya Bashivan
Pablo Samuel Castro
Deep reinforcement learning systems often suffer from unstable training dynamics due to non-stationarity, where learning objectives and data… (voir plus) distributions evolve over time. We show that under non-stationary targets, isotropic Gaussian embeddings are provably advantageous. In particular, they induce stable tracking of time-varying targets for linear readouts, achieve maximal entropy under a fixed variance budget, and encourage a balanced use of all representational dimensions--all of which enable agents to be more adaptive and stable. Building on this insight, we propose the use of Sketched Isotropic Gaussian Regularization for shaping representations toward an isotropic Gaussian distribution during training. We demonstrate empirically, over a variety of domains, that this simple and computationally inexpensive method improves performance under non-stationarity while reducing representation collapse, neuron dormancy, and training instability.
2026-02-21
arXiv (prépublication)
doi.org
arxiv.org