328 citations · 1.6k across the 61 of their papers we have counts for
29 papers · 1 filter
Instruct-SkillMix: A Powerful Pipeline for LLM Instruction Tuning
Simran Kaur, Simon Park, Anirudh Goyal +1
We introduce Instruct-SkillMix, an automated approach for creating diverse, high quality SFT data for instruction-following. The pipeline involves two stages, each leveraging an ex…
Keeping LLMs Aligned After Fine-tuning: The Crucial Role of Prompt Templates
Kaifeng Lyu, Haoyu Zhao, Xinran Gu +3
Public LLMs such as the Llama 2-Chat underwent alignment training and were considered safe. Recently Qi et al. [2024] reported that even benign fine-tuning on seemingly safe datase…
Unlearning via Sparse Representations
Vedant Shah, Frederik Träuble, Ashish Malik +5
Machine \emph{unlearning}, which involves erasing knowledge about a \emph{forget set} from a trained model, can prove to be costly and infeasible by existing techniques. We propose…
A Theory for Emergence of Complex Skills in Language Models
Sanjeev Arora, Anirudh Goyal
A major driver of AI products today is the fact that new skills emerge in language models when their parameter set and training corpora are scaled up. This phenomenon is poorly und…
TC-VAE: Uncovering Out-of-Distribution Data Generative Factors
Cristian Meo, Anirudh Goyal, Justin Dauwels
Uncovering data generative factors is the ultimate goal of disentanglement learning. Although many works proposed disentangling generative models able to uncover the underlying gen…
Discrete Factorial Representations as an Abstraction for Goal Conditioned Reinforcement Learning
Riashat Islam, Hongyu Zang, Anirudh Goyal +6
Goal-conditioned reinforcement learning (RL) is a promising direction for training agents that are capable of solving multiple tasks and reach a diverse set of objectives. How to \…
GFlowOut: Dropout with Generative Flow Networks
Dianbo Liu, Moksh Jain, Bonaventure Dossou +10
Bayesian Inference offers principled tools to tackle many critical problems with modern neural networks such as poor calibration and generalization, and data inefficiency. However,…
Discrete Key-Value Bottleneck
Frederik Träuble, Anirudh Goyal, Nasim Rahaman +4
Deep neural networks perform well on classification tasks where data streams are i.i.d. and labeled data is abundant. Challenges emerge with non-stationary training data streams su…
On the Generalization and Adaption Performance of Causal Models
Nino Scherrer, Anirudh Goyal, Stefan Bauer +2
Learning models that offer robust out-of-distribution generalization and fast adaptation is a key challenge in modern machine learning. Modelling causal structure into neural netwo…
Temporal Latent Bottleneck: Synthesis of Fast and Slow Processing Mechanisms in Sequence Learning
Aniket Didolkar, Kshitij Gupta, Anirudh Goyal +4
Recurrent neural networks have a strong inductive bias towards learning temporally compressed representations, as the entire history of a sequence is represented by a single vector…
Retrieval-Augmented Reinforcement Learning
Anirudh Goyal, Abram L. Friesen, Andrea Banino +13
Most deep reinforcement learning (RL) algorithms distill experience into parametric behavior policies or value functions via gradient updates. While effective, this approach has se…
Variational Causal Networks: Approximate Bayesian Inference over Causal Structures
Yashas Annadani, Jonas Rothfuss, Alexandre Lacoste +4
Learning the causal structure that underlies data is a crucial step towards robust real-world decision making. The majority of existing work in causal inference focuses on determin…
Fast and Slow Learning of Recurrent Independent Mechanisms
Kanika Madan, Nan Rosemary Ke, Anirudh Goyal +2
Decomposing knowledge into interchangeable pieces promises a generalization advantage when there are changes in distribution. A learning agent interacting with its environment is l…
Transformers with Competitive Ensembles of Independent Mechanisms
Alex Lamb, Di He, Anirudh Goyal +4
An important development in deep learning from the earliest MLPs has been a move towards architectures with structural inductive biases which enable the model to keep distinct sour…
Towards Causal Representation Learning
Bernhard Schölkopf, Francesco Locatello, Stefan Bauer +4
The two fields of machine learning and graphical causality arose and developed separately. However, there is now cross-pollination and increasing interest in both fields to benefit…
Inductive Biases for Deep Learning of Higher-Level Cognition
Anirudh Goyal, Yoshua Bengio
A fascinating hypothesis is that human and animal intelligence could be explained by a few principles (rather than an encyclopedic list of heuristics). If that hypothesis was corre…
On Disentangled Representations Learned From Correlated Data
Frederik Träuble, Elliot Creager, Niki Kilbertus +5
The focus of disentanglement approaches has been on identifying independent factors of variation in data. However, the causal variables underlying real-world observations are often…
Diversity inducing Information Bottleneck in Model Ensembles
Samarth Sinha, Homanga Bharadhwaj, Anirudh Goyal +3
Although deep learning models have achieved state-of-the-art performance on a number of vision tasks, generalization over high dimensional multi-modal data, and reliable predictive…
Recurrent Independent Mechanisms
Anirudh Goyal, Alex Lamb, Jordan Hoffmann +4
Learning modular structures which reflect the dynamics of the environment can lead to better generalization and robustness to changes which only affect a few of the underlying caus…
Learning the Arrow of Time
Nasim Rahaman, Steffen Wolf, Anirudh Goyal +2
We humans seem to have an innate understanding of the asymmetric progression of time, which we use to efficiently and safely perceive and manipulate our environment. Drawing inspir…
Reinforcement Learning with Competitive Ensembles of Information-Constrained Primitives
Anirudh Goyal, Shagun Sodhani, Jonathan Binas +3
Reinforcement learning agents that operate in diverse and complex environments can benefit from the structured decomposition of their behavior. Often, this is addressed in the cont…
Learning Powerful Policies by Using Consistent Dynamics Model
Shagun Sodhani, Anirudh Goyal, Tristan Deleu +3
Model-based Reinforcement Learning approaches have the promise of being sample efficient. Much of the progress in learning dynamics models in RL has been made by learning models vi…
State-Reification Networks: Improving Generalization by Modeling the Distribution of Hidden Representations
Alex Lamb, Jonathan Binas, Anirudh Goyal +5
Machine learning promises methods that generalize well from finite labeled data. However, the brittleness of existing neural net approaches is revealed by notable failures, such as…
A Meta-Transfer Objective for Learning to Disentangle Causal Mechanisms
Yoshua Bengio, Tristan Deleu, Nasim Rahaman +5
We propose to meta-learn causal structures based on how fast a learner adapts to new distributions arising from sparse distributional changes, e.g. due to interventions, actions of…