The “Fraction Sense” Emerges from a Deep Convolutional Neural Network

N/A

# link_wmbzbdux30.pdf ## Metadata - PDFFormatVersion=1.5 - IsLinearized=false - IsAcroFormPresent=false - IsXFAPresent=false - IsCollectionPresent=false - IsSignaturesPresent=false - ModDate=D:20200623210324+08'00' - Custom.PTEX.Fullbanner=This is pdfTeX, Version 3.14159265-2.6-1.40.20 (TeX Live 2019) kpathsea version 6.3.1 - CreationDate=D:20200623210324+08'00' - Creator=TeX - Trapped=False - Producer=pdfTeX-1.40.20 ## Contents ### Page 1 The “Fraction Sense” Emerges from a Deep Convolutional Neural NetworkYun-Shiuan Chuang (yunshiuan.chuang@wisc.edu)Department of Psychology, University of

saved by

• Marley Xiong
• Catalina Andrea Saini Ferrón

related reading

• Zoom In: An Introduction to Circuitsdistill.pub
• Toy Models of Superpositiontransformer-circuits.pub
• Neural networks and deep learningneuralnetworksanddeeplearning.com
• Feature Visualizationdistill.pub
• CS231n Deep Learning for Computer Visioncs231n.github.io
• When Models Manipulate Manifolds: The Geometry of a Counting Tasktransformer-circuits.pub
• Convolutional Neural Networks, Explained | Towards Data Sciencetowardsdatascience.com
• Interpretability Dreamstransformer-circuits.pub
• [1811.12231] ImageNet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustnessarxiv.org
• Latest | Epoch AIepochai.org
• Multimodal Neurons in Artificial Neural Networksdistill.pub
• Brain-Score: Which Artificial Neural Network for Object Recognition is most Brain-Like? | bioRxivbiorxiv.org