“I'd recommend working with Bryan any day! He's a stellar data scientist and a go-to person to talk about Math, Graphs and state-of-the-art papers. I especially admire him for not being afraid to ask tough questions. At the same time he's the one who always brings a ton of positive energy, fun and creativity to the room. ”
About
I lead the AI efforts at Hex, where I manage the engineering teams building a Data…
Experience
Education
-
Kansas State University
-
Activities and Societies: Alpha Phi Omega, Philosophy Club, Mathematics Club
Pure mathematics Ph.D. in representation theory and noncommutative algebraic geometry. My thesis is studying representations of quantum groups coming from differential operators on the quantum flag varieties.
-
-
Activities and Societies: Alpha Phi Omega
Started as a Physics/Math/Philosophy major and switched majors 13 times. Finally finished with a math degree and a whole bunch of elective credit after four years.
Volunteer Experience
-
Participated in weekend datadive
DataKind
- 3 months
Human Rights
I participated in a weekend data dive to assist the Santa Clara County Homeless Project. I worked with a team to process, clean, analyze, model, predict, and visualize their data from the past five years. The team produced both forecasts and analysis of the dataset for their use towards improving efficiency and effectiveness..
-
Guest Lecturer
Rutgers University
- 2 months
Education
I gave an hour long lecture on modern forecasting techniques for retail business. I covered the basics of forecasting, metrics for performance evaluation, the application to business, an approach to scaling these models to a large multi-location business, and some of the pitfalls one might run into. The audience was masters students in Analytics and Data Science.
Publications
-
Geometric feature performance under downsampling for EEG classification tasks
We experimentally investigate a collection of feature engineering pipelines for use with a CNN for classifying eyes-open or eyes-closed from electroencephalogram (EEG) time-series from the Bonn dataset. Using the Takens' embedding--a geometric representation of time-series--we construct simplicial complexes from EEG data. We then compare ϵ-series of Betti-numbers and ϵ-series of graph spectra (a novel construction)--two topological invariants of the latent geometry from these complexes--to raw…
We experimentally investigate a collection of feature engineering pipelines for use with a CNN for classifying eyes-open or eyes-closed from electroencephalogram (EEG) time-series from the Bonn dataset. Using the Takens' embedding--a geometric representation of time-series--we construct simplicial complexes from EEG data. We then compare ϵ-series of Betti-numbers and ϵ-series of graph spectra (a novel construction)--two topological invariants of the latent geometry from these complexes--to raw time series of the EEG to fill in a gap in the literature for benchmarking. These methods, inspired by Topological Data Analysis, are used for feature engineering to capture local geometry of the time-series. Additionally, we test these feature pipelines' robustness to downsampling and data reduction. This paper seeks to establish clearer expectations for both time-series classification via geometric features, and how CNNs for time-series respond to data of degraded resolution.
Other authorsSee publication -
Higher order co-occurrence tensors for hypergraphs via face-splitting
A popular trick for computing a pairwise co-occurrence matrix is the product of an incidence matrix and its transpose. We present an analog for higher order tuple co-occurrences using the face-splitting product, or alternately known as the transpose Khatri-Rao product. These higher order co-occurrences encode the commonality of tokens in the company of other tokens, and thus generalize the mutual information commonly studied. We demonstrate this tensor's use via a popular NLP model, and…
A popular trick for computing a pairwise co-occurrence matrix is the product of an incidence matrix and its transpose. We present an analog for higher order tuple co-occurrences using the face-splitting product, or alternately known as the transpose Khatri-Rao product. These higher order co-occurrences encode the commonality of tokens in the company of other tokens, and thus generalize the mutual information commonly studied. We demonstrate this tensor's use via a popular NLP model, and hypergraph models of similarity.
-
The Day It Finally Happens
Simon and Schuster
"From a Vice magazine columnist whose beat is “the future,” here is entertaining speculation featuring both authoritative research and a bit of mischief: a look at how humanity is likely to weather such happenings as the day nuclear war occurs, the day the global internet goes down, the day we run out of effective antibiotics, and the day immortality is achieved." -Simon and Schuster
My colleagues and I consulted in authoring this book, designing the illustrations, and formulating the…"From a Vice magazine columnist whose beat is “the future,” here is entertaining speculation featuring both authoritative research and a bit of mischief: a look at how humanity is likely to weather such happenings as the day nuclear war occurs, the day the global internet goes down, the day we run out of effective antibiotics, and the day immortality is achieved." -Simon and Schuster
My colleagues and I consulted in authoring this book, designing the illustrations, and formulating the research around the technical and data narratives. As consultants we helped in crafting the narrative structures of several chapters, we met with experts across several scientific disciplines, we built mathematical models for prediction and forecasting of some events, and we designed and illustrated the data visualizations that garner many pages of the book.
"The Quasicoherent Labs guys were a godsend, because they offered a service I didn't think existed, but which I nonetheless needed: a one-stop shop that could (as their name suggests) translate my insane, blue-sky ideas into something that makes sense, crunch the relevant numbers, and formulate a visual presentation for the data they found. In my opinion, the illustration that opens the immortality chapter is worth the price of this book alone, because it's so deeply researched and cleverly presented that you can spend a whole morning (as I did) turning it around in your hands, peering at it from different angles, and gaining more and more insight." -Mike Pearl, The day it finally happensOther authorsSee publication -
Integer Coefficient Polynomials have Prime-Rich Images
Mathematics Magazine
Other authors -
On a Basis for the Framed Link Vector Space Spanned by Chord Diagrams
Journal of knot theory and it's ramifications
Courses
-
Introduction to Computational Finance and Financial Econometrics(Coursera)
-
-
Introduction to Data Science(Coursera)
-
Languages
-
French
-
Recommendations received
12 people have recommended Bryan
Join now to viewOther similar profiles
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 MoreOthers named Bryan Bischof
-
Bryan Bischof, MS, RD, CDN
Clinical Nutrition Manager at Norwalk Hospital
-
Bryan Bischof
Shipping/ Document Review/ Control at Humble Machine Works, Inc.
2 others named Bryan Bischof are on LinkedIn
See others named Bryan Bischof