Patrick Shafto
YOU?
Author Swipe
View article: Convergence Theorems for Entropy-Regularized and Distributional Reinforcement Learning
Convergence Theorems for Entropy-Regularized and Distributional Reinforcement Learning Open
In the pursuit of finding an optimal policy, reinforcement learning (RL) methods generally ignore the properties of learned policies apart from their expected return. Thus, even when successful, it is difficult to characterize which polici…
View article: Parent-Child Interaction Styles Relate to Preschooler’s Causal Play, Learning, and Generalization
Parent-Child Interaction Styles Relate to Preschooler’s Causal Play, Learning, and Generalization Open
Parental involvement plays an important role in children’s learning within everyday social contexts. This study investigated whether and how a brief instructional intervention affected parent-child interactions during play, and how parents…
View article: Prosodic Cues Support Inferences About the Question’s Pedagogical Intent
Prosodic Cues Support Inferences About the Question’s Pedagogical Intent Open
Questions may be asked with an intent to acquire new information from the recipient (i.e., information-seeking questions) or with the intent to teach (i.e., pedagogical questions). Understanding how the questions’ recipients infer the inte…
View article: Large Language Models and Children Have Different Learning Trajectories in Determiner Acquisition
Large Language Models and Children Have Different Learning Trajectories in Determiner Acquisition Open
View article: Action Gaps and Advantages in Continuous-Time Distributional Reinforcement Learning
Action Gaps and Advantages in Continuous-Time Distributional Reinforcement Learning Open
When decisions are made at high frequency, traditional reinforcement learning (RL) methods struggle to accurately estimate action values. In turn, their performance is inconsistent and often poor. Whether the performance of distributional …
View article: Young Children’s Directed Question Asking in Preschool Classrooms
Young Children’s Directed Question Asking in Preschool Classrooms Open
Question asking is a prevalent aspect of children’s speech, providing a means by which young learners can rapidly gain information about the world. Previous research has demonstrated that children exhibit sensitivity to the knowledge state…
View article: On Feasibility of Intent Obfuscating Attacks
On Feasibility of Intent Obfuscating Attacks Open
Intent obfuscation is a common tactic in adversarial situations, enabling the attacker to both manipulate the target system and avoid culpability. Surprisingly, it has rarely been implemented in adversarial attacks on machine learning syst…
View article: Structured Evaluation of Synthetic Tabular Data
Structured Evaluation of Synthetic Tabular Data Open
Tabular data is common yet typically incomplete, small in volume, and access-restricted due to privacy concerns. Synthetic data generation offers potential solutions. Many metrics exist for evaluating the quality of synthetic tabular data;…
View article: The time course of adaptation in modified reality: isotropic environments and orientation anisotropies
The time course of adaptation in modified reality: isotropic environments and orientation anisotropies Open
The encoding mechanisms of the human visual system are associated with the distribution of features in natural environments (Olshausen & Field, 2000). Moreover, exposure to modified environments over an hour or so can generate meaningful c…
View article: Coupled Variational Autoencoder
Coupled Variational Autoencoder Open
Variational auto-encoders are powerful probabilistic models in generative tasks but suffer from generating low-quality samples which are caused by the holes in the prior. We propose the Coupled Variational Auto-Encoder (C-VAE), which formu…
View article: Efficient Discretization of Optimal Transport
Efficient Discretization of Optimal Transport Open
Obtaining solutions to optimal transportation (OT) problems is typically intractable when marginal spaces are continuous. Recent research has focused on approximating continuous solutions with discretization methods based on i.i.d. samplin…
View article: Human Variability and the Explore–Exploit Trade‐Off in Recommendation
Human Variability and the Explore–Exploit Trade‐Off in Recommendation Open
The enormous scale of the available information and products on the Internet has necessitated the development of algorithms that intermediate between options and human users. These algorithms attempt to provide the user with relevant infor…
View article: The Inner Loop of Collective Human–Machine Intelligence
The Inner Loop of Collective Human–Machine Intelligence Open
With the rise of artificial intelligence (AI) and the desire to ensure that such machines work well with humans, it is essential for AI systems to actively model their human teammates, a capability referred to as Machine Theory of Mind (MT…
View article: Talk of the Town mobile app platform: New method for engaging family in STEM learning and research in homes and communities
Talk of the Town mobile app platform: New method for engaging family in STEM learning and research in homes and communities Open
Children do not just learn in the classroom. They engage in “informal learning” every day just by spending time with their family and peers. However, while researchers know this occurs, less is known about the science of this learning—how …
View article: The competition–compensation account of developmental language disorder
The competition–compensation account of developmental language disorder Open
Children with developmental language disorder (DLD) regularly use the bare form of verbs (e.g., dance ) instead of inflected forms (e.g., danced ). We propose an account of this behavior in which processing difficulties of children with DL…
View article: Adaptation to the slope of the amplitude spectrum in modified reality
Adaptation to the slope of the amplitude spectrum in modified reality Open
Scenes contain many statistical regularities that could benefit visual processing if accounted for by the visual system. One such statistic to consider is the orientation-averaged slope of the amplitude spectrum of natural scenes. Human ob…
View article: Machine learning modeling practices to support the principles of AI and ethics in nutrition research
Machine learning modeling practices to support the principles of AI and ethics in nutrition research Open
Background Nutrition research is relying more on artificial intelligence and machine learning models to understand, diagnose, predict, and explain data. While artificial intelligence and machine learning models provide powerful modeling to…
View article: Evolution of beliefs in social networks
Evolution of beliefs in social networks Open
Evolution of beliefs of a society are a product of interactions between people (horizontal transmission) in the society over generations (vertical transmission). Researchers have studied both horizontal and vertical transmission separately…
View article: Evaluating perceptual and semantic interpretability of saliency methods: A case study of melanoma
Evaluating perceptual and semantic interpretability of saliency methods: A case study of melanoma Open
In order to be useful, XAI explanations have to be faithful to the AI system they seek to elucidate and also interpretable to the people that engage with them. There exist multiple algorithmic methods for assessing faithfulness, but this i…
View article: Evolution of beliefs in social networks
Evolution of beliefs in social networks Open
Evolution of beliefs of a society are a product of interactions between people (horizontal transmission) in the society over generations (vertical transmission). Researchers have studied both horizontal and vertical transmission separately…
View article: A Psychological Theory of Explainability
A Psychological Theory of Explainability Open
The goal of explainable Artificial Intelligence (XAI) is to generate human-interpretable explanations, but there are no computationally precise theories of how humans interpret AI generated explanations. The lack of theory means that valid…
View article: On Connecting Deep Trigonometric Networks with Deep Gaussian Processes:\n Covariance, Expressivity, and Neural Tangent Kernel
On Connecting Deep Trigonometric Networks with Deep Gaussian Processes:\n Covariance, Expressivity, and Neural Tangent Kernel Open
Deep Gaussian Process (DGP) as a model prior in Bayesian learning intuitively\nexploits the expressive power in function composition. DGPs also offer diverse\nmodeling capabilities, but inference is challenging because marginalization in\n…
View article: On Connecting Deep Trigonometric Networks with Deep Gaussian Processes: Covariance, Expressivity, and Neural Tangent Kernel
On Connecting Deep Trigonometric Networks with Deep Gaussian Processes: Covariance, Expressivity, and Neural Tangent Kernel Open
Deep Gaussian Process (DGP) as a model prior in Bayesian learning intuitively exploits the expressive power in function composition. DGPs also offer diverse modeling capabilities, but inference is challenging because marginalization in lat…
View article: Discrete Probabilistic Inverse Optimal Transport
Discrete Probabilistic Inverse Optimal Transport Open
Optimal transport (OT) formalizes the problem of finding an optimal coupling between probability measures given a cost matrix. The inverse problem of inferring the cost given a coupling is Inverse Optimal Transport (IOT). IOT is less well …
View article: Conditional Deep Gaussian Processes: Multi-Fidelity Kernel Learning
Conditional Deep Gaussian Processes: Multi-Fidelity Kernel Learning Open
Deep Gaussian Processes (DGPs) were proposed as an expressive Bayesian model capable of a mathematically grounded estimation of uncertainty. The expressivity of DPGs results from not only the compositional character but the distribution pr…
View article: Conditional Deep Gaussian Processes: Multi-Fidelity Kernel Learning
Conditional Deep Gaussian Processes: Multi-Fidelity Kernel Learning Open
Deep Gaussian Processes (DGPs) were proposed as an expressive Bayesian model capable of a mathematically grounded estimation of uncertainty. The expressivity of DPGs results from not only the compositional character but the distribution pr…
View article: The Intentional Selection Assumption
The Intentional Selection Assumption Open
There exists a rich literature describing how social context influences decision making. Here, we propose a novel framing of social influences, the Intentional Selection Assumption. This framework proposes that, when a person is presented …
View article: Conditional Deep Gaussian Processes: Empirical Bayes Hyperdata Learning
Conditional Deep Gaussian Processes: Empirical Bayes Hyperdata Learning Open
It is desirable to combine the expressive power of deep learning with Gaussian Process (GP) in one expressive Bayesian learning model. Deep kernel learning showed success as a deep network used for feature extraction. Then, a GP was used a…
View article: Conditional Deep Gaussian Processes: Empirical Bayes Hyperdata Learning
Conditional Deep Gaussian Processes: Empirical Bayes Hyperdata Learning Open
It is desirable to combine the expressive power of deep learning with Gaussian Process (GP) in one expressive Bayesian learning model. Deep kernel learning showed success as a deep network used for feature extraction. Then, a GP was used a…
View article: Abstraction, validation<scp>,</scp> and generalization for explainable artificial intelligence
Abstraction, validation<span>,</span> and generalization for explainable artificial intelligence Open
Neural network architectures are achieving superhuman performance on an expanding range of tasks. To effectively and safely deploy these systems, their decision‐making must be understandable to a wide range of stakeholders. Methods to expl…