In-Context Learning Learns Label Relationships but Is Not Conventional Learning

1 code implementation • 23 Jul 2023 • Jannik Kossen, Yarin Gal, Tom Rainforth

The predictions of Large Language Models (LLMs) on downstream tasks often improve significantly when including examples of the input--label relationship in the context.

In-Context Learning

Three Towers: Flexible Contrastive Learning with Pretrained Image Models

1 code implementation • NeurIPS 2023 • Jannik Kossen, Mark Collier, Basil Mustafa, Xiao Wang, Xiaohua Zhai, Lucas Beyer, Andreas Steiner, Jesse Berent, Rodolphe Jenatton, Efi Kokiopoulou

With 3T, we propose a more flexible strategy that allows the image tower to benefit from both pretrained embeddings and contrastive training.

Contrastive Learning Retrieval

Active Acquisition for Multimodal Temporal Data: A Challenging Decision-Making Task

no code implementations • 9 Nov 2022 • Jannik Kossen, Cătălina Cangea, Eszter Vértes, Andrew Jaegle, Viorica Patraucean, Ira Ktena, Nenad Tomasev, Danielle Belgrave

We introduce a challenging decision-making task that we call active acquisition for multimodal temporal data (A2MT).

Decision Making Informativeness

Marginal and Joint Cross-Entropies & Predictives for Online Bayesian Inference, Active Learning, and Active Sampling

no code implementations • 18 May 2022 • Andreas Kirsch, Jannik Kossen, Yarin Gal

They are more realistic than previously suggested ones, building on work by Wen et al. (2021) and Osband et al. (2022), and focus on evaluating the performance of approximate BNNs in an online supervised setting.

Active Learning Bayesian Inference +1

Active Surrogate Estimators: An Active Learning Approach to Label-Efficient Model Evaluation

1 code implementation • 14 Feb 2022 • Jannik Kossen, Sebastian Farquhar, Yarin Gal, Tom Rainforth

We propose Active Surrogate Estimators (ASEs), a new method for label-efficient model evaluation.

Active Learning

Self-Attention Between Datapoints: Going Beyond Individual Input-Output Pairs in Deep Learning

3 code implementations • NeurIPS 2021 • Jannik Kossen, Neil Band, Clare Lyle, Aidan N. Gomez, Tom Rainforth, Yarin Gal

We challenge a common assumption underlying most supervised deep learning: that a model makes a prediction depending only on its parameters and the features of a single input.

3D Part Segmentation

Active Testing: Sample-Efficient Model Evaluation

1 code implementation • 9 Mar 2021 • Jannik Kossen, Sebastian Farquhar, Yarin Gal, Tom Rainforth

While approaches like active learning reduce the number of labels needed for model training, existing literature largely ignores the cost of labeling test data, typically unrealistically assuming large test sets for model evaluation.

Active Learning Gaussian Processes

Structured Object-Aware Physics Prediction for Video Modeling and Planning

1 code implementation • ICLR 2020 • Jannik Kossen, Karl Stelzner, Marcel Hussing, Claas Voelcker, Kristian Kersting

When humans observe a physical system, they can easily locate objects, understand their interactions, and anticipate future behavior, even in settings with complicated and previously unseen interactions.