Home
deza_MIT_lab_picture.png

Update : Starting August 2023 I will be an Assistant Professor in Computer Science at Universidad de Ingenieria y Tecnologia (UTEC) in Lima, Peru. In parallel, I continue to lead the Artificio team and we will be hiring a new wave of R&D interns in the late Fall of 2023.

I'm currently the Co-Founder and CEO of Artificio, a Moonshot Research & Development for-profit company operating in Cambridge (USA), San Francisco (USA) & Lima (Peru) that aims to solve visual intelligence in humans, machines & everywhere in between. Some of you may have also heard of me after being awarded an EB-1A Green Card for Aliens of Extraordinary Ability in the Sciences to direct my new R&D company after self-petitioning it without a lawyer and though self-study in immigration law. I mostly based my application on the great Razvan Marinescu's open-source LaTeX template.

I was previously a PostDoctoral Research Associate at the McGovern Institute for Brain Research and the Center for Brains, Minds and Machines working with Tomaso Poggio at MIT. Before that, I was a PostDoctoral Research Fellow with Talia Konkle at Harvard University where we worked on understanding the mechanisms of foveation in humans and machines (Pre-print coming soon!)

I completed my PhD in Dynamical Neuroscience at the Vision and Image Understanding Lab at UCSB, where I am supervised by Miguel Eckstein. My previous experience includes working with Manjunath at the Center for BioImage Informatics at UCSB and with Devi Parikh at the Computer Vision and Machine Perception Lab at Virginia Tech. I have a B.S. in Mechatronics Engineering (Robotics) from Universidad Nacional de Ingenieria in Lima, Peru where I was advised by Alberto Coronado and Elizabeth Villota. At UCSB, I was also affiliated with the Institute for Collaborative Biotechnologies.

During my PhD, I worked on towards answering the question of: 'Are biologically inspired human vision models necessary for advanced computer & robot vision? And if so, what aspects should be borrowed and ignored from visual psychology?' More specifically my focus was on understanding the human mechanisms of foveation and peripheral processing, as well as the inverse problem of how advanced computer vision systems may benefit from such peripheral representations. My quest towards answering the two previous questions has led me to perform research at the intersection of vision science, computer vision and machine learning.

You can download my CV here: CV (Last Updated: August 2023) . This personal website is mostly out-dated and you can follow most of my latest research & personal updates in my Google Scholar + Twitter.

You can also find me on Google Scholar, Crunchbase, LinkedIn, GitHub, StackOverflow, Author Influence Graph / Semantic Scholar, Neurotree.

In another life I am an Artist. I was informally trained during long conversations at The Biscuit by John Tarzian. I'll be having a small art show at the Somerville Open Studio in May, 2021. My Art can be purchased here. I am also a proud owner of a Lamborghini

contact: deza {at} artificio . org

Recent & Upcoming Talks:

2021:
Talk at Peru's 3rd Symposium of Deep Learning, January 2021
Talk at Johns Hopkins University's Early Career Colloquium Series, February 2021.
Talk at Coen-Cagli Lab, April 2021.

2020:
Inteview by Modesto Montoya (Spanish)

—-

Talk at the Rosenholtz lab at MIT's CSAIL in Summer of 2017 on Peripheral Representations for Artificial Perception Systems.
Talk at Universidad Nacional de Ingenieria (my alma mater) in Lima, Peru in January 2018 about Peripheral Representations for Artificial Perception Systems.
Talk on Peripheral Representations at the SIAM Applied Mathematics Seminar of UCSB in March 2018.
I will be giving a talk at Redwood Center for Theoretical Neuroscience in March 2018 on Peripheral Representations for computational models of Human and Machine Perception.
Talk at Harvard's Konkle Lab: "Peripheral Representations for Human and Machine Perception" (October 2018)!
Talk at Columbia's Zuckerman Institute: "Probing for Visual Metamerism via Foveated Style Transfer" (October 2018)!
Talk on Visual Metamerism at Google Machine Perception in Mountain View in January 2019!

  • Talk on Exploring the role of Foveation witih Deep Neural Networks at NVIDIA (October 2019), Wolfe lab (Harvard Medical School, February 2020) and Anzellotti Lab (Boston College, April 2020)
  • Talk on Emergent Properties of Foveated Perceptual Systems (MIT Cog Lunch, July 2020)!
  • Talk on Perceptual Invariance as part of the CBMM Virtual Summer School (August, 2020)!
  • Talk on Emergent Properties of Foveated Perceptual Systems (August, 2020) at Facebook Reality Labs!

Recently I have reviewed for / will be reviewing :

  • ICLR 2021
  • Neural Computation 2020/2021
  • NeurIPS 2020
  • ECCV 2020 (Computer Vision, Generative Models)
  • ICML 2020/2021 (Neuroscience, Computer Vision, Deep Learning, Learning Theory)
  • CVPR 2019 & 2020 (Computer Vision)
  • ICCV 2019 (Computer Vision and Representational Learning)
  • CASE 2019 (Robotics),
  • NeurIPS (formerly NIPS) 2019 (Perception, Neuroscience & Deep Learning)
  • CCN 2019 (Cognitive Computational Neuroscience)
  • AAAI 2020

Other news [2020]:

  • I will be TA'ing the relatively new class: "9.58 Projects in the Science of Intelligence" for the new 6-9 course major with Yena Han at MIT in Fall 2020.

Recently received the Harvard Brain Initiative (HBI) Young Scientist Travel Award!

I'm currently co-organizing with Joshua Peterson, Apurva Ratan Murty and Thomas Griffiths, the 1st Shared Visual Representations in Human and Machine Intelligence (SVRHM) Workshop @ NeurIPS 2019.





New Publications Code/Data/Demo
Adversarial_Turtle.png Anne Harrington & Arturo Deza. Finding Biological Plausibility for Adversarially Robust Features via Metameric Tasks. (Spotlight; Top 1% of All Submissions) International Conference on Learning Representations (ICLR), 2022. Code + Data
Gabor_Test.png Chenguang Li & Arturo Deza. What Matters In Branch Specialization? Using a Toy Task to Make Predictions. Shared Visual Representations in Human and Machine Intelligence (SVRHM) workshop at NeurIPS, 2021. Code
Foveated_Texture_Latent.png Jonathan Gant, Andrzej Banburski & Arturo Deza. Evaluating the Adversarial Robustness of a Foveated Texture Transform Module in a CNN. Shared Visual Representations in Human and Machine Intelligence (SVRHM) workshop at NeurIPS, 2021. Code
Foveation_Cat.png Binxu Wang, David Mayo, Arturo Deza, Andrei Barbu & Colin Conwell. On the use of Cortical Magnification and Saccades as Biological Proxies for Data Augmentation. Shared Visual Representations in Human and Machine Intelligence (SVRHM) workshop at NeurIPS, 2021. Code
Pre-Prints Code/Data/Demo
Letters.png Daniel Janini, Chris Hamblin, Arturo Deza, Talia Konkle. General object-based features account for letter perception better than specialized letter features Under Review at Nature Communications, 2021.
Small_Template.png Owen Kunhardt, Arturo Deza*, Tomaso Poggio*. The Effects of Image Distribution and Task on Adversarial Robustness Under Review at ICML, 2021.
Texform_Adjust.png Yi-Chia Chen, Arturo Deza, Talia Konkle. How big should this object be? Perceptual Influences on viewing-size preferences Under Review at Cognition, 2021.
Scrambled_Frog.png Arturo Deza, Qianli Liao, Andrzej Banburski, Tomaso Poggio. Hierarchically Compositional Tasks and Deep Convolutional Networks. ArXiv, 2020. Under Review at ICCV 2021, and currently CBMM Memo #109.
Tweet10_crop.png Arturo Deza, Talia Konkle. Emergent Properties of Foveated Perceptual Systems. ArXiv, 2020. Under review at ICML 2021, and presented at the non-archival venue of VSS (Vision Sciences Society). VSS-Poster VSS Video Poster Walkthrough CBMM tutorial Code + Data
Publications Code/Data/Demo
Cuda_kernel.png Elian Malkin, Arturo Deza*, Tomaso Poggio*. CUDA-Optimized real-time rendering of a Foveated Visual System. Shared Visual Representations in Human and Machine Intelligence (SVRHM) workshop at NeurIPS, 2020. Code Video
640x640_s0.5_a1_o0.5_iter_50.png Arturo Deza, Yi-Chia Chen , Bria Long, Talia Konkle. Accelerated Texforms: Alternative Methods for Generating Unrecognizable Object Images with Preserved Mid-Level Features. Cognitive Computational Neuroscience (CCN), 2019. Code Poster
template_SCORCH_CVPR.png Arturo Deza, Amit Surana, Miguel P. Eckstein. Assessment of Faster R-CNN in Man-Machine collaborative search. IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2019. Coming soon!
Foveated_Storke.png Arturo Deza, Aditya Jonnalagadda, Miguel P. Eckstein. Towards Metamerism via Foveated Style Transfer International Conference on Learning Representations (ICLR), 2019. Lua Code PyTorch Code Poster
Peripheral_Representations.png Arturo Deza. Peripheral Representations: From Perception to Visual Search PhD Thesis , 2018.
d_prime_heatmaps.png Arturo Deza, Jeffrey R. Peters, Grant S. Taylor, Amit Surana, Miguel P. Eckstein. Attention Allocation Aid for Visual Search. Conference on Human Factors in Computing Systems (CHI), 2017. Video
Piranhas_img.png Arturo Deza, Miguel P. Eckstein, Can Peripheral Representations Improve Clutter Metrics on Complex Scenes? Neural Information Processing Systems (NIPS), 2016 NIPS Poster Spotlight Video Video Demo Code
grumpy_cat.jpg Arturo Deza, Devi Parikh, Understanding Image Virality. IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2015. Virality Website
Software
Foveal_Colors.png Arturo Deza, Emre Akbas , Miguel P. Eckstein. Piranhas Toolkit: Peripheral Architectures for Natural, Hybrid and Artificial Systems GitHub
Posters and Workshops
Philosoraptor.png Arturo Deza, Devi Parikh, Understanding Image Virality. IEEE CVPR Scene Understanding Worshop (SUNw) nvidia_pic.png Nvidia Best Poster Award, 2015.
VSS_2015.JPG Arturo Deza, Emre Akbas. , Miguel P. Eckstein, Scene Context and Set Size effects in Search. Vision Sciences Society (VSS) - Visual Search, 2015.
Conference & Symposium Talks (The first author delivered each talk)
AAAD.png Amit Surana, Jeff Peters, Arturo Deza, Grant S. Taylor ,Luca Bertuccelli, Francesco Leonardi , Miguel P. Eckstein. Optimal User Attention Allocation in a Multi-tasking Environment. American Controls Conference (ACC) Workshop, 2016.
148.jpg Arturo Deza, Grant S. Taylor , Miguel P. Eckstein, The Influence of Visual Clutter on Search Guidance with Complex Scenes. Vision Sciences Society (VSS) - Visual Search, 2016
Cosyne.png Miguel P. Eckstein, Arturo Deza,Emre Akbas. Spatial Attention with synthetic cues and real scenes. Cosyne, 2016.
Miscellaneous
cuda_logo.png NeuroSurfer (internet alterego) [Solved] Titan X for CUDA 7.5 login-loop error [Ubuntu 14.04] , Nvidia Forums CUDA Drivers for Linux

News:

  • I joined UTRC (United Technologies Research Center) as a Deep Learning Intern over the summer (2017)!
  • Reviewer of ICCV workshop: https://sites.google.com/site/mbcc2017w/home
  • I will be giving a talk at the Instituto de Matematicas and Ciencias Afines Research Center in Lima, Peru in December on my research.
  • I will be giving a talk at the Mechanical Engineering Department of Universidad Nacional de Ingenieria (my alma mater) in Lima, Peru about my research in December.
  • I will be giving a talk at the Telecommunications Research Center (INICTEL) in Lima, Peru about my research.
  • I am currently serving as an assigned reviewer for IUI 2017.
  • I have server as a volunteer reviewer for NIPS 2016.
  • I am currently organizing the Human and Machine Perception Seminar as part of the Dynamical Neuroscience weekly seminar class.

Ranking of conference proceedings and journals in Artificial Intelligence related fields: Computer Vision, AI and Machine Learning, Robotics, Computer Graphics, Human-Computer Interaction.
Ranking of conference proceedings and journals in Neuroscience related fields: Neurology, Life Sciences

"Give me a computer, a piano, some oils and a surfboard, and I will live happy for the rest of my life" .