Institution profile

INRIA

Academic institutioneurope · fr
Official website
Research library1,897linked papers
Opportunities0open roles
Selected work

Representative Papers

Application-Driven Innovation in Machine Learning

Mar 26, 2024International Conference on Machine Learning

Application-driven machine learning (ML) research has been systematically undervalued in academia, leading to a growing disconnect between algorithmic innovation and real-world needs; this marginalization is reinforced by structural biases in peer review, faculty hiring, and pedagogy. Method: This paper introduces, for the first time, a formally defined “application-driven ML research paradigm,” elucidating its complementary relationship with the dominant methodology-driven paradigm. Drawing on interdisciplinary frameworks from education theory, research governance, and ML practice—and substantiated by empirical case studies and institutional critique—it diagnoses three systemic barriers hindering such research. Contribution/Results: The core contribution is a set of actionable, process-level interventions to reform academic evaluation systems, grounded in both theoretical analysis and pragmatic implementation pathways. These proposals have already catalyzed curricular reforms in AI education and adjustments to national funding review criteria across multiple universities, fostering cross-domain collaboration and methodological feedback loops between application domains and core ML research.

22 citations2 influentialRead paper

What is the Role of Small Models in the LLM Era: A Survey

Sep 10, 2024arXiv.org

Despite the dominance of large language models (LLMs), small models (SMs) remain indispensable in resource-constrained settings due to their low computational cost and deployment flexibility—yet their structural role has been systematically undervalued. Method: This paper introduces the first “collaboration–competition” two-dimensional analytical framework to characterize the dynamic interplay between SMs and LLMs, substantiated through systematic literature review, multi-case empirical comparison, and open-source implementation (GitHub repository). Contribution/Results: The study rigorously delineates SMs’ applicability boundaries across model compression, edge deployment, and human-AI collaboration. It delivers an actionable tripartite guideline covering model selection strategies, lightweight optimization techniques, and deployment paradigms—challenging the misconception that SMs are merely degraded substitutes for LLMs and affirming their foundational complementary role within the AGI ecosystem.

19 citationsRead paper

Querying Inconsistent Prioritized Data with ORBITS: Algorithms, Implementation, and Experiments (Extended Abstract)

Feb 16, 2022Description Logics

This paper addresses inconsistent-tolerant query answering over prioritized knowledge bases—comprising logical theories, factual databases, and priority relations among conflicting facts. We systematically support query evaluation under three classical semantics—AR (cautious), IAR (intersection of all repairs), and brave—over two classes of optimal repair models: Pareto-optimal and completion-based repairs. Our key contribution is the first unified SAT encoding framework capable of handling arbitrary priority relations, enabling joint modeling and efficient solving for both repair classes and all three semantics. Based on this encoding, we implement ORBITS, a novel reasoning system. Experimental results demonstrate that ORBITS significantly outperforms baseline approaches across all semantics, highlighting the critical impact of semantic choice and solving strategy on performance. The work establishes a new paradigm for practical reasoning over inconsistent prioritized knowledge bases.

18 citationsRead paper

Certificates in P and Subquadratic-Time Computation of Radius, Diameter, and all Eccentricities in Graphs

Mar 13, 2018

Computing graph radius, diameter, and all eccentricities is conjectured to require quadratic time under the Strong Exponential Time Hypothesis (SETH). Method: We introduce the notion of *node certificates*—compact auxiliary structures capturing local graph topology sufficient to infer global eccentricities—and establish a tight relationship between certificate size and graph probing complexity. Building on this, we design a randomized subquadratic algorithmic framework supporting queries from one-hop to all-pairs distances, analyzed via primal-dual techniques to yield the first parameterized theoretical guarantees for subquadratic eccentricity computation. Results: Empirical evaluation shows node certificates are significantly smaller than the graph size in real-world networks; our algorithm achieves tight subquadratic time complexity (e.g., $O(n^2 / log n)$) across diverse graph classes; it substantially improves practical runtime while enabling rigorous, provable performance bounds.

11 citationsRead paper

Measure-to-measure interpolation using Transformers

Nov 07, 2024arXiv.org

This paper investigates the expressive power of Transformers as arbitrary input-to-output measure mappings. Method: We reformulate Transformers from a measure-theoretic perspective, modeling them as differentiable maps on continuous measure spaces—departing from conventional discrete token-based interpretations. Leveraging continuity equations to describe particle dynamics, we design an attention mechanism incorporating spherical geometry constraints and optimal transport theory. Contribution/Results: We propose the first Transformer architecture provably capable of exact matching between arbitrary input–target measure pairs. Under the minimal assumption that a transport map exists between each pair, a single model achieves precise matching for N arbitrary measure pairs. We establish theoretical completeness by proving that Transformers serve as universal interpolators between measures and provide explicit parameter constructions. This work fundamentally characterizes the expressive capacity of Transformers for measure transformation tasks.

10 citations1 influentialRead paper
Recent publications

Latest Papers