Deniz Oktay is a scientific researcher and PhD student at Princeton University advised by Ryan P. Adams. They are also a Member of Technical Staff at a startup working on AI for Biology. Oktay has an extensive background in scientific computing and machine learning, with a particular interest in the applications of machine learning to problems in science and engineering.
Education
Oktay completed their BS and MEng in Computer Science at MIT, where they worked with Carl Vondrick and Antonio Torralba.
Experience
Google AI Resident
Intern at Google Brain Princeton with Elad Hazan
Worked at NVIDIA with the Totonto AI Lab
Google X on Project Loon
Yelp working on MOE
Hudson River Trading
Publications
Neuromechanical Autoencoders: Learning to Couple Elastic and Neural Network Nonlinearity (ICLR 2023)
Randomized Automatic Differentiation (ICLR 2021)
Scalable Model Compression by Entropy Penalized Reparameterization (ICLR 2020)
Fiber Monte Carlo (ICLR 2024)
JAX FDM: A differentiable solver for inverse form-finding (ICML 2023)
A rapid and automated computational approach to the design of multistable soft actuators (Computer Physics Communications)
Minuscule corrections to near-surface solar internal rotation using mode-coupling (Astrophysical Journal Supplement Series)
Meta-PDE: Learning to Solve PDEs Quickly Without a Mesh (In submission)
On Predictive Information in RNNs (NeurIPS 2019 Workshop on Information Theory and Machine Learning)
Predicting Motivations of Actions by Leveraging Text (CVPR 2016)
Deniz Oktay
Biography
Deniz Oktay is a PhD student at Princeton University advised by Ryan P. Adams. Oktay is particularly interested in machine learning and its applications in science and engineering. They are currently working as a Member of Technical Staff at a stealth startup focused on AI for Biology.
Education
Oktay has a BS and MEng in Computer Science from MIT, where they worked with Carl Vondrick and Antonio Torralba.
Work Experience
NVIDIA: In Summer 2023, Oktay worked with the Totonto AI Lab, exploring the intersection of numerical simulation and machine learning.
Google Brain Princeton: In Summer 2022, Oktay interned with Elad Hazan.
Google AI Resident: Oktay worked in applied information theory and machine learning.
Hudson River Trading, Google X, and Yelp: Oktay also spent time at these companies, working on various projects.
Publications
Oktay has several notable publications, including:
Neuromechanical Autoencoders: Learning to Couple Elastic and Neural Network Nonlinearity (ICLR 2023):** Co-authored with Mehran Mirramezani, Eder Medina, and Ryan P. Adams.
Randomized Automatic Differentiation (ICLR 2021):** Orally presented at NeurIPS 2020 Beyond Backpropagation Workshop. Co-authored with Nick McGreivy, Joshua Aduol, Alex Beatson, and Ryan P. Adams.
Scalable Model Compression by Entropy Penalized Reparameterization (ICLR 2020):** Co-authored with Johannes Ballé, Saurabh Singh, and Abhinav Shrivastava.
Fiber Monte Carlo (ICLR 2024):** Co-authored with Nick Richardson, Yaniv Ovadia, James C Bowden, and Ryan P. Adams.
JAX FDM: A differentiable solver for inverse form-finding (ICML 2023):** Co-authored with Rafael Pastrana, Ryan P. Adams, and Sigrid Adriaenssens.
Minuscule corrections to near-surface solar internal rotation using mode-coupling(Astrophysical Journal Supplement Series): Co-authored with Srijan Bharati Das, Samarth G. Kashyap, Shravan M. Hanasoge, and Jeroen Tromp.
Meta-PDE: Learning to Solve PDEs Quickly Without a Mesh(In submission): Co-authored with Tian Qin, Alex Beatson, Nick McGreivy, and Ryan P. Adams.
On Predictive Information in RNNs(NeurIPS 2019 Workshop on Information Theory and Machine Learning): Co-authored with Zhe Dong, Ben Poole, and Alex Alemi.
- Predicting Motivations of Actions by Leveraging Text(CVPR 2016): Co-authored with Carl Vondrick, Hamed Pirsiavash, and Antonio Torralba.
Deniz Oktay
Deniz Oktay is a PhD Candidate at Princeton University advised by Ryan P. Adams. Oktay is particularly interested in machine learning and its applications in science and engineering.
Education
Oktay did their BS and MEng in Computer Science at MIT, where they worked with Carl Vondrick and Antonio Torralba.
Work Experience
Oktay has worked at:
NVIDIA with the Totonto AI Lab
Google Brain Princeton as an intern with Elad Hazan
Google as an AI Resident
Hudson River Trading
Google X on Project Loon
Yelp working on MOE
X, the moonshot factory, as a Software Engineer Intern
Tower Research Capital as a Quantitative Trader Intern
Harvard Medical School as an Undergraduate Researcher
Publications
Oktay has worked on the following publications:
Neuromechanical Autoencoders: Learning to Couple Elastic and Neural Network Nonlinearity
Randomized Automatic Differentiation
Scalable Model Compression by Entropy Penalized Reparameterization
Fiber Monte Carlo
JAX FDM: A differentiable solver for inverse form-finding
A rapid and automated computational approach to the design of multistable soft actuators
Minuscule corrections to near-surface solar internal rotation using mode-coupling
Meta-PDE: Learning to Solve PDEs Quickly Without a Mesh
On Predictive Information in RNNs
- Predicting Motivations of Actions by Leveraging Text
Deniz Oktay
Deniz Oktay is a PhD Candidate at Princeton University advised by Ryan P. Adams. Oktay is particularly interested in machine learning and its applications in science and engineering.
Education
Oktay did their BS and MEng in Computer Science at MIT, where they worked with Carl Vondrick and Antonio Torralba.
Work Experience
Oktay has worked at the following companies:
NVIDIA: In Summer 2023, Oktay worked at NVIDIA with the Totonto AI Lab, specifically in the intersection of numerical simulation and machine learning.
Google Brain Princeton: In Summer 2022, Oktay interned at Google Brain Princeton with Elad Hazan.
Google: Oktay was a Google AI Resident, working in the intersection of applied information theory and machine learning.
Hudson River Trading: Oktay worked at Hudson River Trading from January 2018 to January 2019.
Massachusetts Institute of Technology: From January 2017 to January 2018, Oktay was a Graduate Teaching Assistant at MIT.
X, the moonshot factory: From January 2016 to January 2017, Oktay was a Software Engineer Intern at X, the moonshot factory.
Tower Research Capital: From January 2016 to January 2016, Oktay was a Quantitative Trader Intern at Tower Research Capital.
Harvard Medical School: In January 2014, Oktay was an Undergraduate Researcher at Harvard Medical School.
Publications
Oktay has worked on the following publications:
Neuromechanical Autoencoders: Learning to Couple Elastic and Neural Network Nonlinearity: Deniz Oktay, Mehran Mirramezani, Eder Medina, Ryan P. Adams ICLR 2023 (Notable, Top 25% - Spotlight)
Randomized Automatic Differentiation: Deniz Oktay, Nick McGreivy, Joshua Aduol, Alex Beatson, Ryan P. Adams ICLR 2021 (Oral presentation)
Scalable Model Compression by Entropy Penalized Reparameterization: Deniz Oktay, Johannes Ballé, Saurabh Singh, Abhinav Shrivastava ICLR 2020
Fiber Monte Carlo: Nick Richardson, Deniz Oktay, Yaniv Ovadia, James C Bowden, Ryan P. Adams ICLR 2024
JAX FDM: A differentiable solver for inverse form-finding: Rafael Pastrana, Deniz Oktay, Ryan P. Adams, Sigrid Adriaenssens ICML 2023 Differentiable Almost Everything Workshop
A rapid and automated computational approach to the design of multistable soft actuators: Mehran Mirramezani, Deniz Oktay, Ryan P. Adams Computer Physics Communications
Minuscule corrections to near-surface solar internal rotation using mode-coupling: Srijan Bharati Das, Samarth G. Kashyap, Deniz Oktay, Shravan M. Hanasoge, Jeroen Tromp Astrophysical Journal Supplement Series (ApJS)
Meta-PDE: Learning to Solve PDEs Quickly Without a Mesh: Tian Qin, Alex Beatson, Deniz Oktay, Nick McGreivy, Ryan P. Adams In submission
On Predictive Information in RNNs: Zhe Dong, Deniz Oktay, Ben Poole, Alex Alemi NeurIPS 2019 Workshop on Information Theory and Machine Learning
Predicting Motivations of Actions by Leveraging Text: Carl Vondrick, Deniz Oktay, Hamed Pirsiavash, Antonio Torralba CVPR 2016