J. Swaroop Guntupalli
YOU?
Author Swipe
View article: Distributional Diffusion Models with Scoring Rules
Distributional Diffusion Models with Scoring Rules Open
Diffusion models generate high-quality synthetic data. They operate by defining a continuous-time forward process which gradually adds Gaussian noise to data until fully corrupted. The corresponding reverse process progressively "denoises"…
View article: Improving Transformer World Models for Data-Efficient RL
Improving Transformer World Models for Data-Efficient RL Open
We present three improvements to the standard model-based RL paradigm based on transformers: (a) "Dyna with warmup", which trains the policy on real and imaginary data, but only starts using imaginary data after the world model has been su…
View article: Diffusion Model Predictive Control
Diffusion Model Predictive Control Open
We propose Diffusion Model Predictive Control (D-MPC), a novel MPC approach that learns a multi-step action proposal and a multi-step dynamics model, both using diffusion models, and combines them for use in online MPC. On the popular D4RL…
View article: Space is a latent sequence: A theory of the hippocampus
Space is a latent sequence: A theory of the hippocampus Open
Fascinating phenomena such as landmark vector cells and splitter cells are frequently discovered in the hippocampus. Without a unifying principle, each experiment seemingly uncovers new anomalies or coding types. Here, we provide a unifyin…
View article: Graph schemas as abstractions for transfer learning, inference, and planning
Graph schemas as abstractions for transfer learning, inference, and planning Open
Transferring latent structure from one environment or problem to another is a mechanism by which humans and animals generalize with very little data. Inspired by cognitive and neurobiological insights, we propose graph schemas as a mechani…
View article: Space is a latent sequence: Structured sequence learning as a unified theory of representation in the hippocampus
Space is a latent sequence: Structured sequence learning as a unified theory of representation in the hippocampus Open
Fascinating and puzzling phenomena, such as landmark vector cells, splitter cells, and event-specific representations to name a few, are regularly discovered in the hippocampus. Without a unifying principle that can explain these divergent…
View article: Clone-structured graph representations enable flexible learning and vicarious evaluation of cognitive maps
Clone-structured graph representations enable flexible learning and vicarious evaluation of cognitive maps Open
Cognitive maps are mental representations of spatial and conceptual relationships in an environment, and are critical for flexible behavior. To form these abstract maps, the hippocampus has to learn to separate or merge aliased observation…
View article: Hybrid hyperalignment: A single high-dimensional model of shared information embedded in cortical patterns of response and functional connectivity
Hybrid hyperalignment: A single high-dimensional model of shared information embedded in cortical patterns of response and functional connectivity Open
View article: The neural basis of intelligence in fine-grained cortical topographies
The neural basis of intelligence in fine-grained cortical topographies Open
Intelligent thought is the product of efficient neural information processing, which is embedded in fine-grained, topographically organized population responses and supported by fine-grained patterns of connectivity among cortical fields. …
View article: Author response: The neural basis of intelligence in fine-grained cortical topographies
Author response: The neural basis of intelligence in fine-grained cortical topographies Open
Article Figures and data Abstract Introduction Results Discussion Materials and methods Appendix 1 Data availability References Decision letter Author response Article and author information Metrics Abstract Intelligent thought is the prod…
View article: Hybrid Hyperalignment: A single high-dimensional model of shared information embedded in cortical patterns of response and functional connectivity
Hybrid Hyperalignment: A single high-dimensional model of shared information embedded in cortical patterns of response and functional connectivity Open
Shared information content is represented across brains in idiosyncratic functional topographies. Hyperalignment addresses these idiosyncrasies by using neural responses to project individuals’ brain data into a common model space while ma…
View article: From CAPTCHA to Commonsense: How Brain Can Teach Us About Artificial Intelligence
From CAPTCHA to Commonsense: How Brain Can Teach Us About Artificial Intelligence Open
Despite the recent progress in AI powered by deep learning in solving narrow tasks, we are not close to human intelligence in its flexibility, versatility, and efficiency. Efficient learning and effective generalization come from inductive…
View article: The neural basis of intelligence in fine-grained cortical topographies
The neural basis of intelligence in fine-grained cortical topographies Open
Intelligent thought is the product of efficient neural information processing, which is embedded in fine-grained, topographically-organized population responses and supported by fine-grained patterns of connectivity among cortical fields. …
View article: Hyperalignment: Modeling shared information encoded in idiosyncratic cortical topographies
Hyperalignment: Modeling shared information encoded in idiosyncratic cortical topographies Open
Information that is shared across brains is encoded in idiosyncratic fine-scale functional topographies. Hyperalignment captures shared information by projecting pattern vectors for neural responses and connectivities into a common, high-d…
View article: The role of auxiliary parameters in evaluating voxel-wise encoding models for 3T and 7T BOLD fMRI data
The role of auxiliary parameters in evaluating voxel-wise encoding models for 3T and 7T BOLD fMRI data Open
In neuroimaging, voxel-wise encoding models are a popular tool to predict brain activity elicited by a stimulus. To evaluate the accuracy of these predictions across multiple voxels, one can choose between multiple quality metrics. However…
View article: Predicting individual face-selective topography using naturalistic stimuli
Predicting individual face-selective topography using naturalistic stimuli Open
Subject-specific, functionally defined areas are conventionally estimated with functional localizers and a simple contrast analysis between responses to different stimulus categories. Compared with functional localizers, naturalistic stimu…
View article: Memorize-Generalize: An online algorithm for learning higher-order sequential structure with cloned Hidden Markov Models
Memorize-Generalize: An online algorithm for learning higher-order sequential structure with cloned Hidden Markov Models Open
Sequence learning is a vital cognitive function and has been observed in numerous brain areas. Discovering the algorithms underlying sequence learning has been a major endeavour in both neuroscience and machine learning. In earlier work we…
View article: Different clones for different contexts: Hippocampal cognitive maps as higher-order graphs of a cloned HMM
Different clones for different contexts: Hippocampal cognitive maps as higher-order graphs of a cloned HMM Open
Hippocampus encodes cognitive maps that support episodic memories, navigation, and planning. Under-standing the commonality among those maps as well as how those maps are structured, learned from experience, and used for inference and plan…
View article: Memorize-Generalize: An online algorithm for learning higher-order sequential structure with cloned Hidden Markov Models
Memorize-Generalize: An online algorithm for learning higher-order sequential structure with cloned Hidden Markov Models Open
Sequence learning is a vital cognitive function and has been observed in numerous brain areas. Discovering the algorithms underlying sequence learning has been a major endeavour in both neuroscience and machine learning. In earlier work we…
View article: Different clones for different contexts: Hippocampal cognitive maps as higher-order graphs of a cloned HMM
Different clones for different contexts: Hippocampal cognitive maps as higher-order graphs of a cloned HMM Open
View article: Beyond imitation: Zero-shot task transfer on robots by learning concepts as cognitive programs
Beyond imitation: Zero-shot task transfer on robots by learning concepts as cognitive programs Open
Humans can infer concepts from image pairs and apply those in the physical world in a completely different setting, enabling tasks like IKEA assembly from diagrams. If robots could represent and infer high-level concepts, it would signific…
View article: Beyond imitation: Zero-shot task transfer on robots by learning concepts\n as cognitive programs
Beyond imitation: Zero-shot task transfer on robots by learning concepts\n as cognitive programs Open
Humans can infer concepts from image pairs and apply those in the physical\nworld in a completely different setting, enabling tasks like IKEA assembly from\ndiagrams. If robots could represent and infer high-level concepts, it would\nsigni…
View article: Reliable individual differences in fine-grained cortical functional architecture
Reliable individual differences in fine-grained cortical functional architecture Open
Fine-grained functional organization of cortex is not well-conserved across individuals. As a result, individual differences in cortical functional architecture are confounded by topographic idiosyncrasies-i.e., differences in functional-a…
View article: Explaining Visual Cortex Phenomena using Recursive Cortical Network
Explaining Visual Cortex Phenomena using Recursive Cortical Network Open
The connectivity and information pathways of visual cortex are well studied, as are observed physiological phenomena, yet a cohesive model for explaining visual cortex processes remains an open problem. For a comprehensive understanding, w…
View article: A computational model of shared fine-scale structure in the human connectome
A computational model of shared fine-scale structure in the human connectome Open
Variation in cortical connectivity profiles is typically modeled as having a coarse spatial scale parcellated into interconnected brain areas. We created a high-dimensional common model of the human connectome to search for fine-scale stru…
View article: Reliable individual differences in fine-grained cortical functional architecture
Reliable individual differences in fine-grained cortical functional architecture Open
Fine-grained functional organization of cortex is not well-conserved across individuals. As a result, individual differences in cortical functional architecture are confounded by topographic idiosyncrasies—i.e., differences in functional-a…
View article: Cortical Microcircuits from a Generative Vision Model
Cortical Microcircuits from a Generative Vision Model Open
Understanding the information processing roles of cortical circuits is an outstanding problem in neuroscience and artificial intelligence. The theoretical setting of Bayesian inference has been suggested as a framework for understanding co…
View article: The neural representation of personally familiar and unfamiliar faces in the distributed system for face perception
The neural representation of personally familiar and unfamiliar faces in the distributed system for face perception Open
View article: Attention Selectively Reshapes the Geometry of Distributed Semantic Representation
Attention Selectively Reshapes the Geometry of Distributed Semantic Representation Open
Humans prioritize different semantic qualities of a complex stimulus depending on their behavioral goals. These semantic features are encoded in distributed neural populations, yet it is unclear how attention might operate across these dis…
View article: The neural representation of personally familiar and unfamiliar faces in the distributed system for face perception
The neural representation of personally familiar and unfamiliar faces in the distributed system for face perception Open
Personally familiar faces are processed more robustly and efficiently than unfamiliar faces. The human face processing system comprises a core system that analyzes the visual appearance of faces and an extended system for the retrieval of …