Advanced Topics in Deep Learning & IA

Language

English

Prerequisites

Knowledge in machine learning (e.g. LELEC2870, LINFO2262, LDATS2310)

Main themes

The course explores advanced approaches to generative modelling and large-language-model technologies, with a focus on their underlying principles, capabilities, and practical use. The course emphasises hands-on experimentation, critical evaluation, and responsible usage of generative AI systems. The content is updated regularly to reflect rapid technological evolution in the field.

Learning outcomes

At the end of this learning unit, the student is able to :

With regard to the AA framework of the Master [120] in Data Science : Statistic, this activity contributes to the development and acquisition of the following AAs: as first priority: 1.2 – 1.3 – 1.6 – 2.2 – 2.4 – 2.5 – 6.3 as secondary: 3.1 – 3.2 – 5.3 – 5.6 – 6.1 With regard to the AA framework of the Master [120] in Statistics: General, this activity contributes to the development and acquisition of the following AAs: as first priority: 1.3 – 3.2 – 3.3 – 4.1 – 6.3 – 6.4 as secondary: 2.4 – 2.5 – 5.6 – 6.1

Content

• Reminders on deep learning fundamentals.
• Attention mechanisms and Transformer architectures
• LLMs: usage, prompting, inference, fine-tuning
• Generative models (VAEs, GANs).
• Diffusion models.

TPs focus on running local LLMs, prompting methods, and lightweight fine-tuning.

Teaching methods

The course combines Ex-cathedra course supported by slides with practical computer sessions in which students apply each concept using Python notebooks.

Evaluation methods

Oral exam.

Bibliography

“Inside Deep Learning: Math, Algorithms, Models”, Edward Raff, Manning Publications, ISBN-13: 978-1617298639.
“Mastering PyTorch - Second Edition: Create and deploy deep learning models from CNNs to multimodal models, LLMs, and beyond”, Ashish Ranjan Jha, Packt Publishing, ISBN-13: 978-1801074308.

Faculty or entity

> LSBA