Multimodal Foundation Models: time series and beyond ( Mila-Agora + online) - Thu Sept 11, 10am - 1pm
Reasoning in LLMs ( Mila-Agora + online) - Thu Sept 25, 10am - 1pm
LLMs and Psychology ( Mila-Agora + online) - Thu Oct 9, 10am - 1pm
KAIST(Multimodal Foundation Models) ( Mila-Agora + online) - Thu Oct 14, ALL DAY
Scaling and Emergence in New Settings ( Mila-Agora + online) - Fri Oct 25, 10am - 1pm
Scaling Projects Workshop (Mila-Agora/Auditorium2 and online)
AI @ Scale ( Mila-Agora and online, Thu Mar 28 - Fri Mar 29, 2024, follow up Tremblant Mar 30-31)
UdeM/Mila courses:
Fall 2025: Towards AGI: Continual Learning, Scaling and Foundation Models (IFT6167)
Winter 2025: Continual Learning, Scaling and Foundation Models (IFT6167)
Winter 2024: Towards AGI: Scaling, Emergence and Alignment (IFT6760A)
Winter 2023: Towards AGI: Scaling, Alignment and Emergent Behaviors in Neural Nets (IFT6760A)
Winter 2022: Neural Scaling Laws (IFT6760B/6167)
This workshop series are motivated by recent advances in the rapidly developing area of foundation models - large-scale neural networks such as GPT-3, 4 etc - pretrained in an unsupervised way on very large and diverse datasets. Such models often demonstrate significant improvement in their few-shot generalization abilities, as compared to their smaller-scale counterparts, across a wide range of downstream tasks - what one could call a "transformation of quantity into quality" or an "emergent behavior". This is an important step towards a long-standing objective of achieving Artificial General Intelligence (AGI). By AGI here we mean literally a "general", i.e. broad, versatile AI capable of quickly adapting to a wide range of situations and tasks, both novel and those encountered before - i.e. achieving a good stability (memory) vs plasticity (adaptation) trade-off, using the continual learning terminology.
The goal of the workshop series is to provide a forum for discussion of the most recent advances in large-scale pretrained models, focusing specifically on empirical scaling laws of such systems' performance, with increasing compute, model size, and pretraining data (power laws, phase transitions). We will also explore the trade-off between the increasing AI capabilities and AI safety/alignment with human values, considering a range of evaluation metrics beyond the predictive performance. These topics are also closely related to transfer-, continual- and meta-learning, as well as out-of-distribution generalization, robustness and invariant/causal predictive modeling.