Lina Ye

Enseignante-Chercheuse HDR, LMF, CentraleSupélec

Mail: lina.yeREMOVE@lmf.cnrs.fr
Phone: +33 (0)1 75 31 78 69
Office: 3S59, Bâtiment ENS - Nord  

Laboratoire Méthodes Formelles (LMF)
Université Paris-Saclay, CNRS, ENS Paris-Saclay
4 avenue des Sciences • 91190 Gif-sur-Yvette, France

Bio

I studied system information and computer science in University of Surrey, England and University of Paris-Sud 11, France. Then I received the PhD degree in Computer Science from University of Paris-Sud 11 in 2011. I held a post-doctoral position in the research team CONVECS at Inria Rhône-Alpes in 2012-2014. I am currently an associate professor (MCF) of computer science at CentraleSupélec, University Paris-Saclay. I carried out my research work at the LRI laboratory (Laboratoire de Recherche en Informatique) until the end of 2020 and now at the LMF laboratory (Laboratoire Méthodes Formelles), ENS Paris-Saclay, University Paris Saclay, France.

Teaching

I teach the following engineering courses in CentraleSupélec, campus Saclay:

  • algorithmics and complexity (the first year)
  • design and verification of critical system (the second year)
  • intelligent systems for automated air traffic control (the second year)
  • design and verification of hybrid systems (the third year, mention science logiciel)
  • tests (the third year, mention science logiciel)

I also teach the following course for Bachelor AIDAMS of CentraleSupélec and ESSEC, campus Cergy

  • Advanced Programming and Algorithm-Complexity (the second year)

Research

Research Interests

My research interests lie in formal methods for the verification of complex systems, with a focus on both safety and diagnosis-based properties (such as diagnosability, predictability, and manifestability). In particular, my work targets:

  • Infinite-state probabilistic models
  • Machine learning systems (e.g., automata learning algorithm and deep neural networks)
  • Distributed systems
  • Discrete-event, real-time, and hybrid systems

More recently, my research explores interdisciplinary applications, integrating formal methods and machine learning for emotion analysis. One particularly compelling direction is to adapt and develop robust automata learning algorithms (active, passive, and PAC-learning frameworks) to extract discrete emotion states and causal triggers directly from textual interactions. Moving beyond static sentiment classification, this approach maps continuous neural trajectories onto transparent Finite State Transducers (FSTs). Ultimately, these methods scale from text to rich multimodal inputs to reveal structural behavior patterns. The broader objective is to model the cumulative and stateful evolution of human emotions over multi-turn dialogues and clinical transcripts, uncovering how these transitions directly shape human decision-making processes. By replacing opaque black-box representations with inspectable state-transition graphs, researchers and clinicians gain rigorous tools to audit neural models, track evolution pathways, and establish formally verifiable safety invariants for sensitive applications like digital mental health interventions.

Together with colleagues from different disciplines across the University Paris-Saclay perimeter, we organized the first workshop on Computational Psychology with very fruitful feedback that motivates us to continue to work together.

Publications

  • On DBLP.
  • Older publications are also available here.

Current projects

  • PEPR SAIF 2023-2029
  • ANR MAVeriQ 2021-2025

Other activities

  • membre du bureau et membre du conseil pour Guaduate School Informatique et Sciences du Numérique GS-ISN (2021-present)
  • membre du comité des programmes pour l'Institut DATAIA, Université Paris-Saclay (2021-2025)
  • representative of Paris Saclay University for AIDA AIDA - International AI Doctoral Academy (2022-present)