Projects

Software

• Multi-State Markov Model to predict disease evolution in Cancer Research
  Tazi et al. (2022), Nature .

Code

• Neurobiological insights from neural networks trained on facial recognition tasks
  Emergence of Interpretable Functional Specialization in Neural Networks Trained on Facial Expression and Identity Recognition Tazi et al. (2023), ICML WCB.

• Self-Supervised Person-Specific Models
  Objective characterization of an individual’s facial movements using Self-Supervised Learning for Person-Specific-Models Tazi et al. (2022).

• Unsupervised, Statistical Learning and Multi-State Markov Models for Cancer Research
  Unified classification and risk-stratification in Acute Myeloid Leukemia. Tazi et al. (2022), Nature .

• Graph Neural Networks for Drug Discovery
  Graph Neural Network + Attention mechanism to predict scoring functions (i-RMSD) for protein complexes and decoys.

• Multiple Instance Learning for Breast Cancer Metastases
  Multiple Instance Learning with CNN embedding and RNN final classification by information integration using tile embedding.

• Space and Relational Memory with Tolman Eichenbaum Machine
  Building on Whittington et al’s work to evaluate the relationships between hippocampal cells, LEC and MEC, place cells and grid cells across environments using clustering techniques.

• Deep CNNs as a Cloud Service
  Selection and Optimization of Deep Convolutional Neural Networks as a Cloud Service for costs reductions.

• CNN for Breast Cancer and Lesion Classification