Search Results for author: Samuel J. Gershman

Found 19 papers, 4 papers with code

Grokking as the Transition from Lazy to Rich Training Dynamics

no code implementations9 Oct 2023 Tanishq Kumar, Blake Bordelon, Samuel J. Gershman, Cengiz Pehlevan

We identify sufficient statistics for the test loss of such a network, and tracking these over training reveals that grokking arises in this setting when the network first attempts to fit a kernel regression solution with its initial features, followed by late-time feature learning where a generalizing solution is identified after train loss is already low.

Successor-Predecessor Intrinsic Exploration

no code implementations NeurIPS 2023 Changmin Yu, Neil Burgess, Maneesh Sahani, Samuel J. Gershman

Here we focus on exploration with intrinsic rewards, where the agent transiently augments the external rewards with self-generated intrinsic rewards.

Atari Games Efficient Exploration +1

The molecular memory code and synaptic plasticity: a synthesis

no code implementations11 Sep 2022 Samuel J. Gershman

As an alternative, it has been proposed that molecules within the cell body are the storage sites of memory, and that memories are formed through biochemical operations on these molecules.

Representation learning with reward prediction errors

no code implementations27 Aug 2021 William H. Alexander, Samuel J. Gershman

The Reward Prediction Error hypothesis proposes that phasic activity in the midbrain dopaminergic system reflects prediction errors needed for learning in reinforcement learning.

Hippocampus Representation Learning

Human-Level Reinforcement Learning through Theory-Based Modeling, Exploration, and Planning

no code implementations27 Jul 2021 Pedro A. Tsividis, Joao Loula, Jake Burga, Nathan Foss, Andres Campero, Thomas Pouncy, Samuel J. Gershman, Joshua B. Tenenbaum

Here we propose a new approach to this challenge based on a particularly strong form of model-based RL which we call Theory-Based Reinforcement Learning, because it uses human-like intuitive theories -- rich, abstract, causal models of physical objects, intentional agents, and their interactions -- to explore and model an environment, and plan effectively to achieve task goals.

Bayesian Inference Board Games +2

Hybrid Memoised Wake-Sleep: Approximate Inference at the Discrete-Continuous Interface

no code implementations ICLR 2022 Tuan Anh Le, Katherine M. Collins, Luke Hewitt, Kevin Ellis, N. Siddharth, Samuel J. Gershman, Joshua B. Tenenbaum

We build on a recent approach, Memoised Wake-Sleep (MWS), which alleviates part of the problem by memoising discrete variables, and extend it to allow for a principled and effective way to handle continuous variables by learning a separate recognition model used for importance-sampling based approximate inference and marginalization.

Scene Understanding Time Series +1

Analyzing machine-learned representations: A natural language case study

1 code implementation12 Sep 2019 Ishita Dasgupta, Demi Guo, Samuel J. Gershman, Noah D. Goodman

Analyzing performance on these diagnostic tests indicates a lack of systematicity in the representations and decision rules, and reveals a set of heuristic strategies.

What does the free energy principle tell us about the brain?

no code implementations23 Jan 2019 Samuel J. Gershman

The free energy principle has been proposed as a unifying account of brain function.

Active Learning Bayesian Inference

Estimating scale-invariant future in continuous time

no code implementations18 Feb 2018 Zoran Tiganj, Samuel J. Gershman, Per B. Sederberg, Marc W. Howard

Widely used reinforcement learning algorithms discretize continuous time and estimate either transition functions from one step to the next (model-based algorithms) or a scalar value of exponentially-discounted future reward using the Bellman equation (model-free algorithms).

reinforcement-learning Reinforcement Learning (RL)

Evaluating Compositionality in Sentence Embeddings

1 code implementation12 Feb 2018 Ishita Dasgupta, Demi Guo, Andreas Stuhlmüller, Samuel J. Gershman, Noah D. Goodman

Further, we find that augmenting training with our dataset improves test performance on our dataset without loss of performance on the original training dataset.

Natural Language Inference Sentence +2

Deep Successor Reinforcement Learning

1 code implementation8 Jun 2016 Tejas D. Kulkarni, Ardavan Saeedi, Simanta Gautam, Samuel J. Gershman

The successor map represents the expected future state occupancy from any given state and the reward predictor maps states to scalar rewards.

Game of Doom reinforcement-learning +1

Building Machines That Learn and Think Like People

no code implementations1 Apr 2016 Brenden M. Lake, Tomer D. Ullman, Joshua B. Tenenbaum, Samuel J. Gershman

Recent progress in artificial intelligence (AI) has renewed interest in building systems that learn and think like people.

Board Games Object Recognition

Cannot find the paper you are looking for? You can Submit a new open access paper.