Murray Campbell
YOU?
Author Swipe
View article: Fast, slow, and metacognitive thinking in AI
Fast, slow, and metacognitive thinking in AI Open
Inspired by the ”thinking fast and slow” cognitive theory of human decision making, we propose a multi-agent cognitive architecture (SOFAI) that is based on ”fast”/”slow” solvers and a metacognitive module. We then present experimental res…
View article: Position: Theory of Mind Benchmarks are Broken for Large Language Models
Position: Theory of Mind Benchmarks are Broken for Large Language Models Open
Our paper argues that the majority of theory of mind benchmarks are broken because of their inability to directly test how large language models (LLMs) adapt to new partners. This problem stems from the fact that theory of mind benchmarks …
View article: Quantifying artificial intelligence through algorithmic generalization
Quantifying artificial intelligence through algorithmic generalization Open
The rapid development of artificial intelligence (AI) systems has created an urgent need for their scientific quantification. While their fluency across a variety of domains is impressive, AI systems fall short on tests requiring algorithm…
View article: Learning interpretable positional encodings in transformers depends on initialization
Learning interpretable positional encodings in transformers depends on initialization Open
In transformers, the positional encoding (PE) provides essential information that distinguishes the position and order amongst tokens in a sequence. Most prior investigations of PE effects on generalization were tailored to 1D input sequen…
View article: APPROACHES TO AN ACOUSTIC TAXONOMY OF BRASS MUSICAL INSTRUMENTS
APPROACHES TO AN ACOUSTIC TAXONOMY OF BRASS MUSICAL INSTRUMENTS Open
Scholars of brass musical instruments and curators of museums holding collections of historic brasswind have sought to arrange their material by clxsi cation systems.Some (Hombostel & Sachs 1914) and many general writers on musical instntm…
View article: EXPLORER: Exploration-guided Reasoning for Textual Reinforcement Learning
EXPLORER: Exploration-guided Reasoning for Textual Reinforcement Learning Open
Text-based games (TBGs) have emerged as an important collection of NLP tasks, requiring reinforcement learning (RL) agents to combine natural language understanding with reasoning. A key challenge for agents attempting to solve such tasks …
View article: KINETIC ANALYSIS OF HIGH NITROGEN ENERGETIC MATERIALS USING MULTIVARIATE NON-LINEAR REGRESSION
KINETIC ANALYSIS OF HIGH NITROGEN ENERGETIC MATERIALS USING MULTIVARIATE NON-LINEAR REGRESSION Open
New high-nitrogen energetic materials were synthesized by Hiskey and Naud. J. Opfermann reported a new tool for finding the probable model of the complex reactions using multivariate non-linear regression analysis of DSC and TGA data from …
View article: On the generalization capacity of neural networks during generic multimodal reasoning
On the generalization capacity of neural networks during generic multimodal reasoning Open
The advent of the Transformer has led to the development of large language models (LLM), which appear to demonstrate human-like capabilities. To assess the generality of this class of models and a variety of other base neural network archi…
View article: Spectral Enrichment in Valve Trombones
Spectral Enrichment in Valve Trombones Open
Following the development of practicable valves in the 1810s, valved versions of horns, trumpets and trombones quickly followed.The valve trombone became a popular and widely used brass instrument throughout the 19th century, although empl…
View article: The Contributions of Joël Gilbert to the Understanding of `Brassiness
The Contributions of Joël Gilbert to the Understanding of `Brassiness Open
Joël Gilbert was not only an inspired researcher in acoustics but also an excellent and dedicated trombone player. He combined these twin passions in his work on the science of brass musical instruments and was responsible for several impo…
View article: JECC: Commonsense Reasoning Tasks Derived from Interactive Fictions
JECC: Commonsense Reasoning Tasks Derived from Interactive Fictions Open
Commonsense reasoning simulates the human ability to make presumptions about our physical world, and it is an essential cornerstone in building general AI systems. We proposea new commonsense reasoning dataset based on human’s Interactive …
View article: JECC: Commonsense Reasoning Tasks Derived from Interactive Fictions
JECC: Commonsense Reasoning Tasks Derived from Interactive Fictions Open
Commonsense reasoning simulates the human ability to make presumptions about our physical world, and it is an essential cornerstone in building general AI systems. We propose a new commonsense reasoning dataset based on human's Interactive…
View article: Combining Fast and Slow Thinking for Human-like and Efficient Navigation in Constrained Environments
Combining Fast and Slow Thinking for Human-like and Efficient Navigation in Constrained Environments Open
Current AI systems lack several important human capabilities, such as adaptability, generalizability, self-control, consistency, common sense, and causal reasoning. We believe that existing cognitive theories of human decision making, such…
View article: Learning to Teach in Cooperative Multiagent Reinforcement Learning
Learning to Teach in Cooperative Multiagent Reinforcement Learning Open
Collective human knowledge has clearly benefited from the fact that innovations by individuals are taught to others through communication. Similar to human social groups, agents in distributed learning systems would likely benefit from com…
View article: Thinking Fast and Slow in AI: the Role of Metacognition
Thinking Fast and Slow in AI: the Role of Metacognition Open
AI systems have seen dramatic advancement in recent years, bringing many applications that pervade our everyday life. However, we are still mostly seeing instances of narrow AI: many of these recent developments are typically focused on a …
View article: Mental Models of AI Agents in a Cooperative Game Setting (Extended Abstract)
Mental Models of AI Agents in a Cooperative Game Setting (Extended Abstract) Open
As more and more forms of AI become prevalent, it becomes increasingly important to understand how people develop mental models of these systems. In this work we study people's mental models of an AI agent in a cooperative word guessing ga…
View article: Learning to Recover Reasoning Chains for Multi-Hop Question Answering via Cooperative Games
Learning to Recover Reasoning Chains for Multi-Hop Question Answering via Cooperative Games Open
We extend the formats of explanations in interpretable NLP with the proposed entity-centric reasoning chains for multi-hop question answering. We also propose a cooperative game approach to learn to recover such explanations from weakly su…
View article: Text-based RL Agents with Commonsense Knowledge: New Challenges, Environments and Baselines
Text-based RL Agents with Commonsense Knowledge: New Challenges, Environments and Baselines Open
Text-based games have emerged as an important test-bed for Reinforcement Learning (RL) research, requiring RL agents to combine grounded language understanding with sequential decision making. In this paper, we examine the problem of infus…
View article: Circles are like Ellipses, or Ellipses are like Circles? Measuring the Degree of Asymmetry of Static and Contextual Word Embeddings and the Implications to Representation Learning
Circles are like Ellipses, or Ellipses are like Circles? Measuring the Degree of Asymmetry of Static and Contextual Word Embeddings and the Implications to Representation Learning Open
Human judgments of word similarity have been a popular method of evaluating the quality of word embedding. But it fails to measure the geometry properties such as asymmetry. For example, it is more natural to say ``Ellipses are like Circle…
View article: Efficient Text-based Reinforcement Learning by Jointly Leveraging State and Commonsense Graph Representations
Efficient Text-based Reinforcement Learning by Jointly Leveraging State and Commonsense Graph Representations Open
Text-based games (TBGs) have emerged as useful benchmarks for evaluating progress at the intersection of grounded language understanding and reinforcement learning (RL). Recent work has proposed the use of external knowledge to improve the…
View article: Circles are like Ellipses, or Ellipses are like Circles? Measuring the Degree of Asymmetry of Static and Contextual Embeddings and the Implications to Representation Learning
Circles are like Ellipses, or Ellipses are like Circles? Measuring the Degree of Asymmetry of Static and Contextual Embeddings and the Implications to Representation Learning Open
Human judgments of word similarity have been a popular method of evaluating the quality of word embedding. But it fails to measure the geometry properties such as asymmetry. For example, it is more natural to say "Ellipses are like Circles…
View article: Overcoming Catastrophic Forgetting via Direction-Constrained Optimization
Overcoming Catastrophic Forgetting via Direction-Constrained Optimization Open
This paper studies a new design of the optimization algorithm for training deep learning models with a fixed architecture of the classification network in a continual learning framework. The training data is non-stationary and the non-stat…
View article: Continual learning with direction-constrained optimization.
Continual learning with direction-constrained optimization. Open
This paper studies a new design of the optimization algorithm for training deep learning models with a fixed architecture of the classification network in a continual learning framework, where the training data is non-stationary and the no…
View article: Deriving Commonsense Inference Tasks from Interactive Fictions
Deriving Commonsense Inference Tasks from Interactive Fictions Open
Commonsense reasoning simulates the human ability to make presumptions about our physical world, and it is an indispensable cornerstone in building general AI systems. We propose a new commonsense reasoning dataset based on human's interac…
View article: Text-based RL Agents with Commonsense Knowledge: New Challenges, Environments and Baselines
Text-based RL Agents with Commonsense Knowledge: New Challenges, Environments and Baselines Open
Text-based games have emerged as an important test-bed for Reinforcement Learning (RL) research, requiring RL agents to combine grounded language understanding with sequential decision making. In this paper, we examine the problem of infus…
View article: Interactive Fiction Game Playing as Multi-Paragraph Reading Comprehension with Reinforcement Learning
Interactive Fiction Game Playing as Multi-Paragraph Reading Comprehension with Reinforcement Learning Open
Interactive Fiction (IF) games with real human-written natural language texts provide a new natural evaluation for language understanding techniques. In contrast to previous text games with mostly synthetic texts, IF games pose language un…
View article: A Study of Compositional Generalization in Neural Models
A Study of Compositional Generalization in Neural Models Open
Compositional and relational learning is a hallmark of human intelligence, but one which presents challenges for neural models. One difficulty in the development of such models is the lack of benchmarks with clear compositional and relatio…