“Greg was an outstanding asset to the Data Science team at Uptake. When Greg first joined our team, he quickly made an impact within our R&D, focusing primarily on satellite imagery. He then went on to make other contributions to our core supervised learning engine, developing a new approach to model evaluation that led to a patent application. Greg makes for a great teammate because he is meticulous, transparent, and organized and is great at both receiving and giving feedback in productive ways. Not only that, Greg volunteered his time and made contributions to the Data Science community by teaching others both internally and externally to Uptake. Greg would be a fantastic addition to any data science team and I would happily work with him again.”
About
I am a data scientist modeling AV safety at Cruise. I have previously developed machine…
Activity
-
I want to express my deepest gratitude to the Motion Planning & Prediction teams at Argo AI. Over the past year, we landed deep learning-powered…
I want to express my deepest gratitude to the Motion Planning & Prediction teams at Argo AI. Over the past year, we landed deep learning-powered…
Liked by Greg Gandenberger, PhD
-
Many fantastic people from Argo AI are looking for new opportunities. Please consider them if you are hiring! https://lnkd.in/gYNzyMUY
Many fantastic people from Argo AI are looking for new opportunities. Please consider them if you are hiring! https://lnkd.in/gYNzyMUY
Shared by Greg Gandenberger, PhD
Experience
Education
-
University of Pittsburgh
-
- Developed a new proof of the Likelihood Principle and discussed its implications for debates about Bayesian, frequentist, and likelihoodist paradigms within statistics.
- Published articles in Philosophy of Science and the British Journal for the Philosophy of Science, widely recognized as the two top journals in the field.
- Gave multiple conference presentations.
- Taught courses on scientific reasoning. -
-
- Admitted into cohort of 61 out of 3021 applicants
- Tools used include Postregres, SQL, Hadoop MapReduce, Spark, NumPy, SciPy, pandas, scikit-learn, librosa, NetworkX, matplotlib, Bokeh, D3, and Flask.
- Completed nine weekly projects on topics such as social network analysis, time series modeling, and distributed computing.
- Deployed a Pitchf/x data visualization flask app.
- Used Pitchf/x data and traditional statistics to predict the ERAs of Major League Baseball free agent…- Admitted into cohort of 61 out of 3021 applicants
- Tools used include Postregres, SQL, Hadoop MapReduce, Spark, NumPy, SciPy, pandas, scikit-learn, librosa, NetworkX, matplotlib, Bokeh, D3, and Flask.
- Completed nine weekly projects on topics such as social network analysis, time series modeling, and distributed computing.
- Deployed a Pitchf/x data visualization flask app.
- Used Pitchf/x data and traditional statistics to predict the ERAs of Major League Baseball free agent pitchers with greater accuracy than standard sabermetric approaches. -
-
Studied experimental design, generalized linear models, ANOVA, nonparametric statistics, likelihood theory, measure theory, Markov processes, Brownian motion.
-
-
Activities and Societies: Phi Beta Kappa
- Wrote honor's thesis that was published in the journal Philosophy of Science.
- Completed 24 semester-hours in physics and 21 in math in addition to philosophy major.
Licenses & Certifications
Volunteer Experience
-
First LEGO League Robotics Coach / Mentor
Lincoln Middle School
- Present 1 year 10 months
Education
Overseeing two First LEGO League Robotics teams at Lincoln Middle School
-
Data Scientist
Lincoln Park Zoo Urban Wildlife Institute
- 2 years 5 months
Environment
Trained and deployed a model that identifies animals in pictures from motion-activated "camera traps." Researchers use those images to monitor wildlife populations. The model removes a major bottleneck in their work by reducing the number of images that they need to label manually.
"It was like they gifted us 10 interns who could work around the clock," said Urban Wildlife Institute researcher Mason Fidino. "Autofocus enables us to get on top of our backlog of images and deliver valuable…Trained and deployed a model that identifies animals in pictures from motion-activated "camera traps." Researchers use those images to monitor wildlife populations. The model removes a major bottleneck in their work by reducing the number of images that they need to label manually.
"It was like they gifted us 10 interns who could work around the clock," said Urban Wildlife Institute researcher Mason Fidino. "Autofocus enables us to get on top of our backlog of images and deliver valuable insights on local wildlife now." -
Data Fellows Mentor
beyond.uptake
- Present 7 years 4 months
Poverty Alleviation
- Gave presentations on tidy data, the grammar of graphics, and overhead imagery
- Mentored Hello Tractor data scientist -
Volunteer
Uptake
- Present 7 years 5 months
Education
Built benches for John W. Cook Elementary School library as part of Uptake Day of Service.
-
Volunteer
Long Beach Church of Christ
- 4 months
Disaster and Humanitarian Relief
Helped residents with Hurricane Katrina recovery through activities such as roofing, painting and trash cleanup.
Publications
-
Differences Among Noninformative Stopping Rules Are Often Relevant to Bayesian Decisions
Arxiv
L.J. Savage once hoped to show that "the superficially incompatible systems of ideas associated on the one hand with [subjective Bayesianism] and on the other hand with [classical statistics]...lend each other mutual support and clarification." By 1972, however, he had largely "lost faith in the devices" of classical statistics. One aspect of those "devices" that he found objectionable is that differences among the "stopping rules" that are used to decide when to end an experiment which are…
L.J. Savage once hoped to show that "the superficially incompatible systems of ideas associated on the one hand with [subjective Bayesianism] and on the other hand with [classical statistics]...lend each other mutual support and clarification." By 1972, however, he had largely "lost faith in the devices" of classical statistics. One aspect of those "devices" that he found objectionable is that differences among the "stopping rules" that are used to decide when to end an experiment which are "noninformative" from a Bayesian perspective can affect decisions made using a classical approach. Two experiments that produce the same data using different stopping rules seem to differ only in the intentions of the experimenters regarding whether or not they would have carried on if the data had been different, which seem irrelevant to the evidential import of the data and thus to facts about what actions the data warrant.
I argue that classical and Bayesian ideas about stopping rules do in fact "lend each other" the kind of "mutual support and clarification" that Savage had originally hoped to find. They do so in a kind of case that is common in scientific practice, in which those who design an experiment have different interests from those who will make decisions in light of its results. I show that, in cases of this kind, Bayesian principles provide qualified support for the classical statistical practice of "penalizing" "biased" stopping rules. However, they require this practice in a narrower range of circumstances than classical principles do, and for different reasons. I argue that classical arguments for this practice are compelling in precisely the class of cases in which Bayesian principles also require it, and thus that we should regard Bayesian principles as clarifying classical statistical ideas about stopping rules rather than the reverse. -
Why I Am Not a Likelihoodist
Philosophers' Imprint
Frequentist statistical methods continue to predominate in many areas of science despite prominent calls for "statistical reform." They do so in part because their main rivals, Bayesian methods, appeal to prior probability distributions that arguably lack an objective justification in typical cases. Some methodologists find a third approach called likelihoodism attractive because it avoids important objections to frequentism without appealing to prior probabilities. However, likelihoodist…
Frequentist statistical methods continue to predominate in many areas of science despite prominent calls for "statistical reform." They do so in part because their main rivals, Bayesian methods, appeal to prior probability distributions that arguably lack an objective justification in typical cases. Some methodologists find a third approach called likelihoodism attractive because it avoids important objections to frequentism without appealing to prior probabilities. However, likelihoodist methods do not provide guidance for belief or action, but only assessments of data as evidence. I argue that there is no good way to use those assessments to guide beliefs or actions without appealing to prior probabilities, and that as a result likelihoodism is not a viable alternative to frequentism and Bayesianism for statistical reform efforts in science.
-
A New Proof of the Likelihood Principle
The British Journal for the Philosophy of Science
I present a new proof of the likelihood principle that avoids two responses to a well-known proof due to Birnbaum ([1962]). I also respond to arguments that Birnbaum’s proof is fallacious, which if correct could be adapted to this new proof. On the other hand, I urge caution in interpreting proofs of the likelihood principle as arguments against the use of frequentist statistical methods.
-
Teaching Domain-General Skills Throughout the Undergraduate Curriculum
University of Pittsburgh Teaching Times
-
Producing a Robust Body of Data with a Single Technique
Philosophy of Science
When a technique purports to provide information that is not available to the unaided senses, it is natural to think that the only way to validate that technique is by appealing to a theory of the processes that lead from the object of study to the raw data. In fact, scientists have a variety of strategies for validating their techniques. Those strategies can yield multiple independent arguments that support the validity of the technique. Thus, it is possible to produce a robust body of data…
When a technique purports to provide information that is not available to the unaided senses, it is natural to think that the only way to validate that technique is by appealing to a theory of the processes that lead from the object of study to the raw data. In fact, scientists have a variety of strategies for validating their techniques. Those strategies can yield multiple independent arguments that support the validity of the technique. Thus, it is possible to produce a robust body of data with a single technique. I illustrate and support these claims with a historical case study.
Patents
-
Computer system and method for evaluating an event prediction model
Issued US 11119472
If A and B are two event prediction models where A produces both more true positives ("catches") and more false positives ("false flags") than B, choose A over B if the ratio of the number of additional catches from A to the number of additional false flags from A is above a fixed ratio that reflects the relative costs of false negatives to false positives.
-
Computer system and method for classifying temporal patterns of change in images of an area
Issued US 10255526
System for automatically identifying temporal patterns of change (e.g. transient, persistent, or reverting), for instance in satellite images.
Courses
-
Advanced Logic
2500
-
Advanced Philosophy of Science
4210
-
Applied Nonparametric Statistics
2200
-
Applied Statistical Methods 1 & 2
2131, 2132
-
Biostatistics Seminar
2025
-
Calculus III
233
-
Causality
2660
-
Cognition and Knowledge
2630
-
Determinism
2642
-
Differential Equations
217
-
Electricity and Magnetism I & II
421, 422
-
Evidence and Risk (focused on assessment of Evidence-Based Medicine movement)
2622
-
Foundations for Higher Mathematics
310
-
Intermediate Mathematical Statistics
1632
-
Intermediate Probability
2630
-
Introduction and Relational Databases (Stanford Lagunita)
-
-
Introduction to Computational Thinking and Data Science (MITx)
6.00.2x
-
Introduction to Computer Science and Programming Using Python (MITx)
6.00.1x
-
Introduction to Machine Learning for Coders (fast.ai)
-
-
Introduction to Quantum Physics I & II
217, 318
-
Introduction to Relativity
216
-
Likelihood Theory and Applications
2061
-
Linear Algebra 1
1180
-
Logic and Computation (Carnegie Mellon University)
0610
-
Machine Learning (Andrew Ng's Stanford course materials)
229
-
Mathematical Statistics
494
-
Ockham's Razor (Carnegie Mellon University)
0524
-
Part 2: Deep Learning from the Foundations (fast.ai)
-
-
Philosophy of Biological Science
423
-
Philosophy of Science
321G
-
Philosophy of Science Core Seminar
2501
-
Physics I & II
197, 198
-
Physics and Philosophy (St. Louis University)
417
-
Practical Deep Learning for Coders (fast.ai)
-
-
Probability
493
-
Probability Theory 1 & 2
2711, 2712
-
Pyquick (Google)
-
-
Python (Codecademy)
-
-
Relational Algebra (Stanford Lagunita)
-
-
SQL (Stanford Lagunita)
-
-
SQLZoo
-
-
Seminar on Foundations of Statistics (Carnegie Mellon University)
0515
-
Statistical Mechanics and Thermodynamics
463
-
The Gene: Transformation and Fragmentation of a Concept
2565
-
The SQL Tutorial for Data Analysis (Mode)
-
-
Theories of Confirmation
2682
-
Theory of Statistics 1
2631
-
Topics in Applied Statistics 2: Clinical Trial Regulations
3132
Projects
-
How to Think about Bayesianism, Frequentism, and Likelihoodism
Uptake Tech Talk
-
Causal Inference from Big Data
Uptake journal club presentation
-
Overhead Imagery Analytics
Presented to beyond.uptake data fellows
-
A Layered Grammar of Graphics
Uptake journal club presentation
-
Tidy Data
Uptake journal club presentation
-
Why Reproducibility Matters
Presentation for Reproducibility Project Workshop, Cardiff University, Cardiff, Wales.
-
Statistical Inference: Formal Epistemology Meets Scientific Practice.
Presentation at Formal Epistemology Summer School, University of Bristol, Bristol, UK
-
A Bayesian Vindication of the Frequentist Position on Stopping Rules.
Presentation in Work in Progress Series, Munich Center for Mathematical Philosophy, Ludwig Maximilians University, Munich, Germany
-
Significance Testing and the Likelihood Principle
Guest lecture in Philosophy of Statistics course, Munich Center for Mathematical Philosophy, Ludwig Maximilians University, Munich, Germany
-
New Responses to Purported Counterexamples to Likelihoodist Principles
Presentation for Philosophy of Science Colloquium, Munich Center for Mathematical Philosophy, Ludwig Maximilians University, Munich, Germany
-
Why It Is Sometimes Permissible to Violate the Likelihood Principle—Even If It Is True
Presentation for Workshop on Inductive Logic and Confirmation in Science II, University of Utah, Salt Lake City, UT
-
The Trouble with Likelihoodism
Presented at History and Philosophy of Science Seminar, Washington University in St. Louis, St. Louis, MO
-
Why the Law of Likelihood Applies Only to Mutually Exclusive Hypotheses
Presentation at American Philosophical Association Pacific Division Meeting, San Diego, CA
-
A New Proof of the Likelihood Principle
Presented at Philosophy of Science Association Biennial Meeting, San Diego, CA
-
Why I Am Not a Methodological Likelihoodist
-
Presented at Philosophy of Science Association Biennial Meeting, Chicago, IL and at Meeting of the Society for Exact Philosophy, California Institute of Technology, Pasadena, CA
Honors & Awards
-
Graduate Student Fellowship
Dolores Zohrab Liebman Fund
Nationally competitive; 13 awarded.
-
Graduate Student Fellowship
Josephine de Karman Fellowship Trust
Nationally competitive; 9 recipients were selected out of more than 400 applicants.
-
Outstanding Presentation Prize
University of Pittsburgh Graduate Student Organization
-
Elizabeth Baranger Teaching Award
University of Pittsburgh Graduate Student Organization
-
Nisha Luthra Prize
Washington University in St. Louis Department of Philosophy
Awarded to the top graduating student in philosophy, as chosen by the department faculty.
-
Beinecke Scholarship
The Sperry Fund
Nationally competitive; 22 awarded.
-
Student Research Grant
Washington University in St. Louis Center for the Study of Ethics and Human Values
Funded archival research on the introduction of the cathode-ray oscillograph into electrophysiology
Test Scores
-
GRE
Score: 800 / 800 / 5.5
-
SAT
Score: 800 / 800
Organizations
-
Philosophy of Science Association
-
- -
American Philosophical Association
-
Recommendations received
1 person has recommended Greg
Join now to viewMore activity by Greg
-
Hey, Germany! If you see us in Munich, give us a wave. We’ve been working on our #selfdriving technology, logging miles to transform how people live…
Hey, Germany! If you see us in Munich, give us a wave. We’ve been working on our #selfdriving technology, logging miles to transform how people live…
Liked by Greg Gandenberger, PhD
-
UX Research, fundamentally, is about empathy and advocacy. My team spent several weeks listening to and learning from customers in the gender…
UX Research, fundamentally, is about empathy and advocacy. My team spent several weeks listening to and learning from customers in the gender…
Liked by Greg Gandenberger, PhD
-
I was pretty inspired by #dalle and #Imagen so I grabbed the open source version made by Boris Dayma 🥑 and created https://www.visual-if.com .…
I was pretty inspired by #dalle and #Imagen so I grabbed the open source version made by Boris Dayma 🥑 and created https://www.visual-if.com .…
Liked by Greg Gandenberger, PhD
-
I have been at SpotHero for less than two months but have really enjoyed the great atmosphere and potential for growth. We have lots of great…
I have been at SpotHero for less than two months but have really enjoyed the great atmosphere and potential for growth. We have lots of great…
Liked by Greg Gandenberger, PhD
-
I'm having a ton of fun here. I'm learning so much from people who know way more than I do. Check it out if you're lookin'!
I'm having a ton of fun here. I'm learning so much from people who know way more than I do. Check it out if you're lookin'!
Liked by Greg Gandenberger, PhD
People also viewed
Explore collaborative articles
We’re unlocking community knowledge in a new way. Experts add insights directly into each article, started with the help of AI.
Explore More