Isaac V Kauvar

ikauvar@


> about


Hello! I am an LSRF Postdoctoral Fellow at Stanford University, working with Nick Haber in the Stanford Autonomous Agents Lab at the intersection of artificial intelligence and neuroscience. Broadly, I am interested in how networks of neurons yield intelligence and behavior, and how we can apply this knowledge for the betterment of the world. In practice, I study the links between neural network dynamics and behaviors such as exploration, self-protection, and goal-directed decision making. The sources of data I work with include multiregion cellular recording in mice, intracranial EEG in humans, and AI systems (e.g. model-based deep reinforcement learning agents). For more detail on one of my current projects, in which I am building bridges between AI and animal physiology to improve AI exploration and elucidate the neural basis of curiosity, see here, here, and here.

I earned my PhD at Stanford University, studying neuroscience with Karl Deisseroth and Gordon Wetzstein, and supported by an NSF Graduate Research Fellowship. There, I developed tools for recording and analyzing cortex-wide neuronal activity and discovered a neural oscillation—surprisingly localized to one cortical region—that appears to underlie dissociative effects of drugs such as ketamine.


> publications


Policy-shaped prediction: avoiding distractions in model-based RL.
Hutson M, Kauvar I, Haber N.
NeurIPS (2024).
[PDF]
Summary: Applies gradient-based interpretability to improve the robustness of model-based RL agents against distractions.
Curious Replay for Model-based Adaptation.
Kauvar I* , Doyle CR*, Zhou L, Haber N. * = equal contribution
ICML (2023).
[PDF][arxiv][Website][Code][Article]Hacker News]
Summary: Inspired by animal behavior, a method that improves model-based RL adaptivity and achieves a new state-of-the-art score on Crafter.
Interaction modeling with multiplex attention.
Sun FY, Kauvar I , Zhang R, Li J, Kochenderfer M, Wu J, Haber N.
NeurIPS (2022).
[PDF][link]
Summary: A method for modeling the dynamics of multi-agent systems, and achieving state-of-the-art trajectory forecasting and relation inference.
Multiregion neuronal activity: the forest and the trees.
Machado TA*, Kauvar I* , Deisseroth K. * = equal contribution
Nature Reviews Neuroscience (2022). (cover article)
[PDF][link]
Summary: A survey of tools for recording and analyzing brainwide neuronal activity, and insights from seeing a simultaneously broad and detailed view.
Deep posteromedial cortical rhythm in dissociation.
Vesuna S*, Kauvar I*, Richman E, Gore F, Oskotsky T, Sava-Segal C, Luo L, Malenka RC, Henderson JM, Nuyujukian P, Parvizi J, Deisseroth K. * = equal contribution
Nature (2020).
[PDF][link][News&Views][Research Highlight][Stanford press][NPR radio segment][HackerNews][The Independent][TikTok][Knowing Neurons]
Summary: Brainwide recording reveals surprisingly localized rhythmic activity linked to ketamine and dissociation in mice and humans.
Cortical observation by synchronous multifocal optical sampling reveals widespread population encoding of actions.
Kauvar I*, Machado T*, Yuen E, Kochalka J, Choi M, Allen WE, Wetzstein G, Deisseroth K. * = equal contribution
Neuron (2020).
[site: use COSMOS!][PDF][supplement][link][press]
Summary: A new technique (COSMOS) for recording fast cortex-wide neuronal activity reveals multiregion encoding of current and future actions.
A light-field metasurface for high-resolution single-particle tracking.
Holsteen A*, Lin D*, Kauvar I, Wetzstein G, Brongersma M. * = equal contribution
Nano Letters (2019). [PDF][link]
Ancestral circuits for the coordinated modulation of brain state.
Lovett-Barron M, Andalman AS, Allen WE, Vesuna S, Kauvar I, Burns VM, Deisseroth K.
Cell (2017). [PDF]
Modulation of prefrontal cortex excitation/inhibition balance rescues social behavior in CNTNAP2-deficient mice.
Selimbeyoglu A, Kim CK, Inoue M, Lee SY, Hong ASO, Kauvar I, Ramakrishnan C, Fenno LE, Davidson TJ, Wright M, Deisseroth K.
Science Trans. Med. (2017). [PDF]
Global representations of goal-directed behavior in distinct cell types of mouse neocortex.
Allen WE*, Kauvar I*, Chen MZ, Richman E, Yang SJ, Chan K, Gradinaru V, Deverman BE, Luo L, Deisseroth K.
Neuron (2017). * = equal contribution [link][PDF][press]
Aperture interference and the volumetric resolution of light field fluorescence microscopy.
Kauvar I, Chang J, Wetzstein G.
IEEE Intl. Conf. on Comp. Photog. (ICCP) (2017). [site][PDF][supplement]
Variable Aperture Light Field Photography: Overcoming the Diffraction-limited Spatio-angular Resolution Tradeoff.
Chang J, Kauvar I, Hu X, Wetzstein G.
IEEE CVPR (2016). [site][PDF][supplement]
Simultaneous fast measurement of circuit dynamics at multiple sites across the mammalian brain.
Kim CK*, Yang SJ*, Pichamoorthy N, Young NP, Kauvar I, Jennings JH, Lerner TN, Berndt A, Lee SY, Ramakrishnan C, Davidson TJ, Inoue M, Bito H, Deisseroth K. * = equal contribution
Nature Methods (2016). [PDF][supplement]
SPED Light Sheet Microscopy: Fast Mapping of Biological System Structure and Function.
Tomer R, Lovett-Barron M, Kauvar I, Andalman AS, Burns VM, Sankaran S, Grosenick L, Broxton M, Yang S, Deisseroth K.
Cell 163:7 (2015). [PDF]
Extended field-of-view and increased-signal 3D holographic illumination with time-division multiplexing.
Yang SJ, Allen WE, Kauvar I, Andalman AS, Young NP, Kim CK, Marshel JH, Wetzstein G., Deisseroth K.
Optics Express 23:25 (2015). [link] [site] [PDF]
Adaptive color display via perceptually-driven factored spectral projection.
Kauvar I, Yang SJ, Shi, L, McDowall I, Wetzstein G.
ACM Transactions on Graphics (TOG) (2015): 165. [PDF]
Natural Neural Projection Dynamics Underlying Social Behavior.
Gunaydin LA*, Grosenick L*, Finkelstein JC*, Kauvar IV*, Fenno LE, Adhikari A, Lammel S, Mirzabekov JJ, Airan RD, Zalocusky KA, Tye KM, Anikeeva P, Malenka RC, Deisseroth K. * = equal contribution
Cell (2014). [PDF ][ Stanford Medicine][Scope Blog][LA Times][TIME]
The Role of Electron Affinity in Determining Whether Fullerenes Catalyze or Inhibit Photooxidation of Polymers for Solar Cells.
Hoke E, Sachs Quintana IT, Lloyd MT, Kauvar IV, Mateker WR, Nardes AM, Peters CH, Kopidakis N, McGehee MD.
Advanced Energy Materials (2012). [PDF]
Unconventional Face-On Texture and Exceptional In-Plane Order of a High Mobility n-Type Polymer.
Rivnay J, Toney MF, Zheng Y, Kauvar IV, Chen Z, Wagner V, Facchetti A, Salleo A.
Advanced Materials (2010). [PDF]


> patents

Multifocal macroscope for large field of view imaging of dynamic specimens
Gordon Wetzstein, Tim Machado, Karl A. Deisseroth, Isaac V. Kauvar. US20210364771A1
Method and systems for measuring neural activity
Christina Kim, Samuel Yang, Karl A. Deisseroth, Isaac V. Kauvar. US11147457B2
Methods and devices for imaging and/or optogenetic control of light-responsive neurons
Samuel Yang, Karl A. Deisseroth, William E. Allen, Isaac V. Kauvar, Aaron S. Andalman. US10568516B2
Liquid crystal displays with reduced light leakage
Yi Huang, Pongpinit Towashiraporn, Erik A. Zweigle, Mingxia Gu, Wei Chen, Cheng Chen, Young Cheol Yang, Isaac V. Kauvar. US9081230B2

> art


I like to draw; I design clothing apparel you can buy here!
I am a photographer; check out some of my work here and here. I have also published a book of elephant portraits, Bigger Than Your Self, available for purchase on Amazon here .
I also enjoy making movies; I helped produce a film Carrying Capacity about research into the neuroscience of autism. It won Best Short Documentary at the 2023 Vail Film Festival.



> previous


In 2019 I worked on an ambitious 'moonshot' using machine learning (mostly Graph Neural Networks) as an AI resident at Google X, Google's secretive R&D lab.

In 2014 I worked as a computational imaging engineer at Light.

I spent the summer of 2013 traveling along the northern Mediterranean, shooting rolls and rolls of black and white film. These photographs led to a solo exhibition in the Stanford SubGallery (many of which can be viewed here).

In 2012-2013 I completed a M.Sc. in Electrical Engineering at Stanford University on an IEEE Charles Legeyt Fortescue Scholarship. I studied primarily machine learning, image processing, and photography.

During the summer of 2012, I worked at Apple as part of the Panel Process and Optics team where I leveraged my background in optical physics to improve the liquid crystal display stack, resulting in this patent.

Before I happened upon imaging, photography, and neuroscience, I was most passionate about renewable (in particular, solar) energy conversion. As an undergraduate at Stanford University, majoring in Engineering Physics, I worked with Alberto Salleo and at the National Renewable Energy Laboratory, yielding two publications in the realm of improved organic semiconductors and lifetime stability of organic photovoltaics.



> things I have built


COSMOS
Record neural activity at video-rates across the extent of rodent dorsal cortex, with near-single-cell resolution.

Optoencephalography
Record neural activity with mesoscale resolution in genetically-defined cell-types across rodent dorsal cortex.

Multi-fiber photometry
Record neural activity in populations of genetically- or projection-defined cell-types simultaneously, from many regions (cortical and subcortical) throughout the brain.

Ferrofluid music visualizer
An iPod controlled sound system with a dynamic equalizer and a ferrofluid wave chamber controlled by electromagnets synced to the music. (The linked video shows the not-quite-finished product).