AIConfPaper
← Search

Michael Curtis Mozer

21 accepted papers

2026

Redirection for Erasing Memory (REM): Towards a universal unlearning method for corrupted data

Stefan Schoepf, Michael Curtis Mozer, Nicole Elyse Mitchell, Alexandra Brintrup, Georgios Kaissis, Peter Kairouz +1
ICLR 2026poster

Machine unlearning is studied for a multitude of tasks, but specialization of unlearning methods to particular tasks has made their systematic comparison challenging. To address this issue, we propose a conceptual space to characterize diverse corrupted data unlearning tasks in vision classifiers. T…

Cited by 0Source ↗Scholar ↗
2025

From Dormant to Deleted: Tamper-Resistant Unlearning Through Weight-Space Regularization

Shoaib Ahmed Siddiqui, Adrian Weller, David Krueger, Gintare Karolina Dziugaite, Michael Curtis Mozer, Eleni Triantafillou
NeurIPS 2025poster

Recent unlearning methods for LLMs are vulnerable to relearning attacks: knowledge believed-to-be-unlearned re-emerges by fine-tuning on a small set of (even seemingly-unrelated) examples. We study this phenomenon in a controlled setting for example-level unlearning in vision classifiers. We make th…

Cited by 0Source ↗Scholar ↗
2025

On the Transfer of Object-Centric Representation Learning

Aniket Rajiv Didolkar, Andrii Zadaianchuk, Anirudh Goyal, Michael Curtis Mozer, Yoshua Bengio, Georg Martius +1
ICLR 2025poster

The goal of object-centric representation learning is to decompose visual scenes into a structured representation that isolates the entities into individual vectors. Recent successes have shown that object-centric representation learning can be scaled to real-world scenes by utilizing features from…

Cited by 1Source ↗PDF ↗Scholar ↗
2025

Racing Thoughts: Explaining Contextualization Errors in Large Language Models

Michael A. Lepori, Michael Curtis Mozer, Asma Ghandeharioun
NAACL 2025long

The profound success of transformer-based language models can largely be attributed to their ability to integrate relevant contextual information from an input sequence in order to generate a response or complete a task. However, we know very little about the algorithms that a model employs to imple…

Cited by 0Source ↗PDF ↗Scholar ↗
2024

Reawakening knowledge: Anticipatory recovery from catastrophic interference via structured training

Yanlai Yang, Matt Jones, Michael Curtis Mozer, Mengye Ren
NeurIPS 2024poster

We explore the training dynamics of neural networks in a structured non-IID setting where documents are presented cyclically in a fixed, repeated sequence. Typically, networks suffer from catastrophic interference when training on a sequence of documents; however, we discover a curious and remarkabl…

Cited by 1Source ↗PDF ↗code ↗Scholar ↗
2024

Recurrent Complex-Weighted Autoencoders for Unsupervised Object Discovery

Anand Gopalakrishnan, Aleksandar Stanić, Jürgen Schmidhuber, Michael Curtis Mozer
NeurIPS 2024poster

Current state-of-the-art synchrony-based models encode object bindings with complex-valued activations and compute with real-valued weights in feedforward architectures. We argue for the computational advantages of a recurrent architecture with complex-valued weights. We propose a fully convolutiona…

Cited by 2Source ↗PDF ↗code ↗Scholar ↗
2023

Can Neural Network Memorization Be Localized?

Pratyush Maini, Michael Curtis Mozer, Hanie Sedghi, Zachary Chase Lipton, J Zico Kolter, Chiyuan Zhang
ICML 2023poster

Recent efforts at explaining the interplay of memorization and generalization in deep overparametrized networks have posited that neural networks *memorize* ``hard'' examples in the final few layers of the model. Memorization refers to the ability to correctly predict on *atypical* examples of the t…

Cited by 59Source ↗PDF ↗code ↗Scholar ↗
2023

Discrete Key-Value Bottleneck

Frederik Träuble, Anirudh Goyal, Nasim Rahaman, Michael Curtis Mozer, Kenji Kawaguchi, Yoshua Bengio +1
ICML 2023poster

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 such as continual learning. One powerful approach that has addressed this challenge involves pre-training of large encoder…

Cited by 30Source ↗PDF ↗code ↗Scholar ↗
2023

Learning in temporally structured environments

Matt Jones, Tyler R. Scott, Mengye Ren, Gamaleldin Fathy Elsayed, Katherine Hermann, David Mayo +1
ICLR 2023poster

Natural environments have temporal structure at multiple timescales. This property is reflected in biological learning and memory but typically not in machine learning systems. We advance a multiscale learning method in which each weight in a neural network is decomposed as a sum of subweights with…

Cited by 6Source ↗PDF ↗Scholar ↗
2023

Learning to Induce Causal Structure

Nan Rosemary Ke, Silvia Chiappa, Jane X Wang, Jorg Bornschein, Anirudh Goyal, Melanie Rey +4
ICLR 2023poster

The fundamental challenge in causal induction is to infer the underlying graph structure given observational and/or interventional data. Most existing causal induction algorithms operate by generating candidate graphs and evaluating them using either score-based methods (including continuous optimiz…

Cited by 59Source ↗PDF ↗Scholar ↗
2023

Stateful Active Facilitator: Coordination and Environmental Heterogeneity in Cooperative Multi-Agent Reinforcement Learning

Dianbo Liu, Vedant Shah, Oussama Boussif, Cristian Meo, Anirudh Goyal, Tianmin Shu +3
ICLR 2023poster

In cooperative multi-agent reinforcement learning, a team of agents works together to achieve a common goal. Different environments or tasks may require varying degrees of coordination among agents in order to achieve the goal in an optimal way. The nature of coordination will depend on properties o…

Cited by 10Source ↗PDF ↗Scholar ↗
2022

Coordination Among Neural Modules Through a Shared Global Workspace

Anirudh Goyal, Aniket Rajiv Didolkar, Alex Lamb, Kartikeya Badola, Nan Rosemary Ke, Nasim Rahaman +4
ICLR 2022oral

Deep learning has seen a movement away from representing examples with a monolithic hidden state towards a richly structured state. For example, Transformers segment by position, and object-centric architectures decompose images into entities. In all these architectures, interactions between differe…

Cited by 108Source ↗PDF ↗Scholar ↗
2022

SAVi++: Towards End-to-End Object-Centric Learning from Real-World Videos

Gamaleldin Fathy Elsayed, Aravindh Mahendran, Sjoerd van Steenkiste, Klaus Greff, Michael Curtis Mozer, Thomas Kipf
NeurIPS 2022accept

The visual world can be parsimoniously characterized in terms of distinct entities with sparse interactions. Discovering this compositional structure in dynamic visual scenes has proven challenging for end-to-end computer vision approaches unless explicit instance-level supervision is provided. Slot…

Cited by 154Source ↗PDF ↗Scholar ↗
2021

Discrete-Valued Neural Communication

Dianbo Liu, Alex Lamb, Kenji Kawaguchi, Anirudh Goyal, Chen Sun, Michael Curtis Mozer +1
NeurIPS 2021poster

Deep learning has advanced from fully connected architectures to structured models organized into components, e.g., the transformer composed of positional elements, modular architectures divided into slots, and graph neural nets made up of nodes. The nature of structured models is that communication…

Cited by 58Source ↗PDF ↗Scholar ↗
2021

Factorizing Declarative and Procedural Knowledge in Structured, Dynamical Environments

Anirudh Goyal, Alex Lamb, Phanideep Gampa, Philippe Beaudoin, Charles Blundell, Sergey Levine +2
ICLR 2021poster

Modeling a structured, dynamic environment like a video game requires keeping track of the objects and their states (declarative knowledge) as well as predicting how objects behave (procedural knowledge). Black-box models with a monolithic hidden state often fail to apply procedural knowledge consis…

Cited by 16Source ↗PDF ↗Scholar ↗
2021

Improving Anytime Prediction with Parallel Cascaded Networks and a Temporal-Difference Loss

Michael Louis Iuzzolino, Michael Curtis Mozer, Samy Bengio
NeurIPS 2021poster

Although deep feedforward neural networks share some characteristics with the primate visual system, a key distinction is their dynamics. Deep nets typically operate in serial stages wherein each layer completes its computation before processing begins in subsequent layers. In contrast, biologica…

Cited by 17Source ↗PDF ↗code ↗Scholar ↗
2021

Neural Production Systems

Anirudh Goyal, Aniket Rajiv Didolkar, Nan Rosemary Ke, Charles Blundell, Philippe Beaudoin, Nicolas Heess +2
NeurIPS 2021poster

Visual environments are structured, consisting of distinct objects or entities. These entities have properties---visible or latent---that determine the manner in which they interact with one another. To partition images into entities, deep-learning researchers have proposed structural inductive bia…

Cited by 96Source ↗PDF ↗Scholar ↗
2021

Soft Calibration Objectives for Neural Networks

Archit Karandikar, Nicholas Cain, Dustin Tran, Balaji Lakshminarayanan, Jonathon Shlens, Michael Curtis Mozer +1
NeurIPS 2021poster

Optimal decision making requires that classifiers produce uncertainty estimates consistent with their empirical accuracy. However, deep neural networks are often under- or over-confident in their predictions. Consequently, methods have been developed to improve the calibration of their predictive un…

Cited by 97Source ↗PDF ↗Scholar ↗
2021

Systematic Evaluation of Causal Discovery in Visual Model Based Reinforcement Learning

Nan Rosemary Ke, Aniket Rajiv Didolkar, Sarthak Mittal, Anirudh Goyal, Guillaume Lajoie, Stefan Bauer +4
NeurIPS 2021poster

Inducing causal relationships from observations is a classic problem in machine learning. Most work in causality starts from the premise that the causal variables themselves are observed. However, for AI agents such as robots trying to make sense of their environment, the only observables are low-l…

Cited by 52Source ↗code ↗Scholar ↗
2021

Wandering within a world: Online contextualized few-shot learning

Mengye Ren, Michael Louis Iuzzolino, Michael Curtis Mozer, Richard Zemel
ICLR 2021poster

We aim to bridge the gap between typical human and machine-learning environments by extending the standard framework of few-shot learning to an online, continual setting. In this setting, episodes do not have separate training and testing phases, and instead models are evaluated online while learnin…

Cited by 41Source ↗PDF ↗code ↗Scholar ↗