no code implementations • 2 May 2024 • Parikshit Ram, Tim Klinger, Alexander G. Gray
We then show how various existing general and special purpose sequence processing models (such as recurrent, convolution and attention-based ones) fit this definition and use it to analyze their compositional complexity.
1 code implementation • 15 Mar 2024 • Kinjal Basu, Keerthiram Murugesan, Subhajit Chaudhury, Murray Campbell, Kartik Talamadupula, Tim Klinger
To tackle these issues, in this paper, we present EXPLORER which is an exploration-guided reasoning agent for textual reinforcement learning.
1 code implementation • 28 Sep 2023 • Tim Klinger, Luke Liu, Soham Dan, Maxwell Crouse, Parikshit Ram, Alexander Gray
Compositional generalization is a key ability of humans that enables us to learn new concepts from only a handful examples.
no code implementations • 7 May 2023 • Maxwell Crouse, Pavan Kapanipathi, Subhajit Chaudhury, Tahira Naseem, Ramon Astudillo, Achille Fokoue, Tim Klinger
Nearly all general-purpose neural semantic parsers generate logical forms in a strictly top-down autoregressive fashion.
no code implementations • 30 Mar 2023 • Tyler Malloy, Miao Liu, Matthew D. Riemer, Tim Klinger, Gerald Tesauro, Chris R. Sims
This raises the question of how humans learn to efficiently represent visual information in a manner useful for learning tasks.
no code implementations • 15 Sep 2022 • Takuya Ito, Tim Klinger, Douglas H. Schultz, John D. Murray, Michael W. Cole, Mattia Rigotti
Our findings give empirical support to the role of compositional generalization in human behavior, implicate abstract representations as its neural implementation, and illustrate that these representations can be embedded into ANNs by designing simple and efficient pretraining procedures.
1 code implementation • 1 Mar 2022 • JunKyu Lee, Michael Katz, Don Joven Agravante, Miao Liu, Geraud Nangue Tasse, Tim Klinger, Shirin Sohrabi
Our approach defines options in hierarchical reinforcement learning (HRL) from AIP operators by establishing a correspondence between the state transition model of AI planning problem and the abstract state transition system of a Markov Decision Process (MDP).
no code implementations • AAAI Workshop CLeaR 2022 • Kinjal Basu, Keerthiram Murugesan, Mattia Atzeni, Pavan Kapanipathi, Kartik Talamadupula, Tim Klinger, Murray Campbell, Mrinmaya Sachan, Gopal Gupta
These rules are learned in an online manner and applied with an ASP solver to predict an action for the agent.
Inductive logic programming Natural Language Understanding +2
no code implementations • 23 Nov 2020 • Tyler Malloy, Tim Klinger, Miao Liu, Matthew Riemer, Gerald Tesauro, Chris R. Sims
This paper introduces an information-theoretic constraint on learned policy complexity in the Multi-Agent Deep Deterministic Policy Gradient (MADDPG) reinforcement learning algorithm.
no code implementations • 9 Oct 2020 • Tyler Malloy, Chris R. Sims, Tim Klinger, Miao Liu, Matthew Riemer, Gerald Tesauro
We focus on the model-free reinforcement learning (RL) setting and formalize our approach in terms of an information-theoretic constraint on the complexity of learned policies.
no code implementations • 16 Jun 2020 • Tim Klinger, Dhaval Adjodah, Vincent Marois, Josh Joseph, Matthew Riemer, Alex 'Sandy' Pentland, Murray Campbell
One difficulty in the development of such models is the lack of benchmarks with clear compositional and relational task structure on which to systematically evaluate them.
2 code implementations • 28 Apr 2020 • Cameron Allen, Michael Katz, Tim Klinger, George Konidaris, Matthew Riemer, Gerald Tesauro
Focused macros dramatically improve black-box planning efficiency across a wide range of planning domains, sometimes beating even state-of-the-art planners with access to a full domain model.
no code implementations • 25 Sep 2019 • Tyler James Malloy, Matthew Riemer, Miao Liu, Tim Klinger, Gerald Tesauro, Chris R. Sims
We formalize this type of bounded rationality in terms of an information-theoretic constraint on the complexity of policies that agents seek to learn.
no code implementations • NAACL 2019 • Ignacio Cases, Clemens Rosenbaum, Matthew Riemer, Atticus Geiger, Tim Klinger, Alex Tamkin, Olivia Li, S Agarwal, hini, Joshua D. Greene, Dan Jurafsky, Christopher Potts, Lauri Karttunen
The model jointly optimizes the parameters of the functions and the meta-learner{'}s policy for routing inputs through those functions.
1 code implementation • 29 Apr 2019 • Clemens Rosenbaum, Ignacio Cases, Matthew Riemer, Tim Klinger
Compositionality is a key strategy for addressing combinatorial complexity and the curse of dimensionality.
no code implementations • 6 Sep 2018 • Andres Campero, Aldo Pareja, Tim Klinger, Josh Tenenbaum, Sebastian Riedel
Our approach is neuro-symbolic in the sense that the rule pred- icates and core facts are given dense vector representations.
no code implementations • 17 Nov 2017 • Matthew Riemer, Tim Klinger, Djallel Bouneffouf, Michele Franceschini
Given the recent success of Deep Learning applied to a variety of single tasks, it is natural to consider more human-realistic settings.
1 code implementation • ICLR 2018 • Shuohang Wang, Mo Yu, Jing Jiang, Wei zhang, Xiaoxiao Guo, Shiyu Chang, Zhiguo Wang, Tim Klinger, Gerald Tesauro, Murray Campbell
We propose two methods, namely, strength-based re-ranking and coverage-based re-ranking, to make use of the aggregated evidence from different passages to better determine the answer.
Ranked #1 on Open-Domain Question Answering on Quasar
no code implementations • ICLR 2018 • Clemens Rosenbaum, Tim Klinger, Matthew Riemer
A routing network is a kind of self-organizing neural network consisting of two components: a router and a set of one or more function blocks.
1 code implementation • 31 Aug 2017 • Shuohang Wang, Mo Yu, Xiaoxiao Guo, Zhiguo Wang, Tim Klinger, Wei zhang, Shiyu Chang, Gerald Tesauro, Bo-Wen Zhou, Jing Jiang
Second, we propose a novel method that jointly trains the Ranker along with an answer-generation Reader model, based on reinforcement learning.
Ranked #4 on Open-Domain Question Answering on Quasar
no code implementations • 5 Aug 2017 • Clemens Rosenbaum, Tian Gao, Tim Klinger
In this paper we present a new dataset and user simulator e-QRAQ (explainable Query, Reason, and Answer Question) which tests an Agent's ability to read an ambiguous text; ask questions until it can answer a challenge question; and explain the reasoning behind its questions and answer.
4 code implementations • 2 Jun 2016 • Iulian Vlad Serban, Tim Klinger, Gerald Tesauro, Kartik Talamadupula, Bo-Wen Zhou, Yoshua Bengio, Aaron Courville
We introduce the multiresolution recurrent neural network, which extends the sequence-to-sequence framework to model natural language generation as two parallel discrete stochastic processes: a sequence of high-level coarse tokens, and a sequence of natural language tokens.
Ranked #1 on Dialogue Generation on Ubuntu Dialogue (Activity)