( events)   Timezone: »  
Mon Dec 13 06:45 AM -- 03:00 PM (PST)
Shared Visual Representations in Human and Machine Intelligence
Arturo Deza · Joshua Peterson · N Apurva Ratan Murty · Tom Griffiths

Workshop Home Page

The goal of the 3rd Shared Visual Representations in Human and Machine Intelligence \textit{(SVRHM)} workshop is to disseminate relevant, parallel findings in the fields of computational neuroscience, psychology, and cognitive science that may inform modern machine learning. In the past few years, machine learning methods---especially deep neural networks---have widely permeated the vision science, cognitive science, and neuroscience communities. As a result, scientific modeling in these fields has greatly benefited, producing a swath of potentially critical new insights into the human mind. Since human performance remains the gold standard for many tasks, these cross-disciplinary insights and analytical tools may point towards solutions to many of the current problems that machine learning researchers face (\textit{e.g.,} adversarial attacks, compression, continual learning, and self-supervised learning). Thus we propose to invite leading cognitive scientists with strong computational backgrounds to disseminate their findings to the machine learning community with the hope of closing the loop by nourishing new ideas and creating cross-disciplinary collaborations. In particular, this year's version of the workshop will have a heavy focus on testing new inductive biases on novel datasets as we work on tasks that go beyond object recognition.

Opening Remarks (Remarks)
Closing Remarks + Award Presentation (Remarks)
Bio-inspired learnable divisive normalization for ANNs (Poster)
Multimodal neural networks better explain multivoxel patterns in the hippocampus (Oral)
Combining Different V1 Brain Model Variants to Improve Robustness to Image Corruptions in CNNs (Poster)
In Silico Modelling of Neurodegeneration Using Deep Convolutional Neural Networks (Poster)
Shared Visual Representations of Drawing for Communication: How do different biases affect human interpretability and intent? (Poster)
Boxhead: A Dataset for Learning Hierarchical Representations (Poster)
Convolutional Networks are Inherently Foveated (Poster)
Out-of-distribution robustness: Limited image exposure of a four-year-old is enough to outperform ResNet-50 (Poster)
A finer mapping of convolutional neural network layers to the visual cortex (Poster)
Benchmarking human visual search computational models in natural scenes: models comparison and reference datasets (Poster)
Controlled-rearing studies of newborn chicks and deep neural networks (Oral)
Learning to perceive objects by prediction (Oral)
KDSalBox: A toolbox of efficient knowledge-distilled saliency models (Poster)
What Matters In Branch Specialization? Using a Toy Task to Make Predictions (Poster)
Finding Biological Plausibility for Adversarially Robust Features via Metameric Tasks (Poster)
Evaluating the Adversarial Robustness of a Foveated Texture Transform Module in a CNN (Poster)
Cyclic orthogonal convolutions for long-range integration of features (Poster)
Bio-inspired Min-Nets Improve the Performance and Robustness of Deep Networks (Poster)
Exploiting 3D Shape Bias towards Robust Vision (Poster)
Modeling Category-Selective Cortical Regions with Topographic Variational Autoencoders (Oral)
What can 5.17 billion regression fits tell us about artificial models of the human visual system? (Poster)
Are models trained on temporally-continuous data streams more adversarially robust? (Poster)
Neural Structure Mapping For Learning Abstract Visual Analogies (Poster)
Signal Strength and Noise Drive Feature Preference in CNN Image Classifiers (Poster)
V1 and IT representations are directly accessible to human visual perception (Poster)
Unsupervised Representation Learning Facilitates Human-like Spatial Reasoning (Poster)
Exploring the Structure of Human Adjective Representations (Poster)
Contrastive Learning Through Time (Poster)
On the use of Cortical Magnification and Saccades as Biological Proxies for Data Augmentation (Poster)
Category-orthogonal object features guide information processing in recurrent neural networks trained for object categorization (Poster)
Seeking the Building Blocks of Visual Imagery and Creativity in a Cognitively Inspired Neural Network (Poster)