M Pacer 📨

I am a core contributor to the Jupyter Project.

I am a Senior Notebook Engineer at Netflix.

In 2016, I graduated from UC Berkeley as a member of Tania Lombrozo's Cognition and Concepts Laboratory and Tom Griffith's Computational Cognitive Science Laboratory.

I am a computational cognitive scientist — I make maths on minds.

Below is my research programme from my PhD, last updated long ago.

Analog Clock12:00:00 AM

    If you can see this message, your browser does not support canvas and needs an update. Sorry. :(

Programme(in progress):

Updated: 2016_06_5_0333

time
- temporal causation
- ordering and approximate ordinality
- mixed-class stochastic processes
- continuous-time & discrete-time systems interacting
causal measure
- extending Noisy-OR/AND-NOT
- as analogs of generation and prevention
- to both continuous measures and continuous-time
- (possibly) as exemplified by causal information flow
calculus of ideas
- generalization of causal information flow to continuous-time stochastic processes, over continuous dimensions and discrete/structured representations.
  - measure chains have analysis, and may be an appropriate tool.
continuous theories
- generalization gradients
- similarity spaces
- compositionality
- commeasurability of data despite embedding in different spaces

explanatory virtues
- simplicity
- reaism
- notions of probability
formal models
- explanation as antibody
- explanation as orgasm (from Gopnik's Explanation as Orgasm and the drive for causal understanding)
- causal explantory trees & causal information flow
- Generalized Bayes Factor
science of the particular

formal models of cultural learning
conductive bias
- the role of notational systems in the biases transmitted over time
iterative learning
serial reproduction
science of the artificial
small samples & neutral bias currents
- see also, [continuity][]
history of
- mathematical/logical notation
- writing systems
- formal logic [Pierce — Gödel]
- science
- philosophy of science
- medicine
- epidemiology
- nosology
- anatomy as a categorization problem
- notions of continuity
- notions of contagion
- notions of mechanism
- publishing/typesetting
- statistics
- mathematics
- biology & philosophy of biology
- taxonomy as categorization problem
- phenetics/cladistics
- units & levels of selection
- design
  - experimental
- architecture
- typographic (& other graphic design)
design
- experimental
- architecture
- typographic (& other graphic design)
  - letter-spacing
- serifs
- communication-design (written and otherwise)

Levels of Analysis
- rational analysis
- extending Marr's levels
Hierarchical Bayesian networks
- causal Bayesian networks
- continuous-time Bayesian networks
- Poisson cascades
- Markov Modulated poisson process
- theory-based causal induction
Probabilistic, compositional semantics (a newer interest)
- e.g., the Church language
- caching/memoizing
Topic models & other
- dynamic topic models
- discrete time
- continuous time
- extensions that allow only a finite sub-dictionary to be attached to a topic
- tags
- unsupervised strict taxonomization
- dictionary learning
- sense disambiguation & polysemy
Non parametric Bayseian Inference
- Chinese restaurant process (mutually exclusive categories)
- Indian Buffet process (additive featural representations)
- Transformations over non-parametric processes
Inferential methods
- Variational Inference
- Stochastic variational inference
- optimization theory as applies to VI
- Monte Carlo (sampling) methods
- Markov Chain Monte Carlo
- Hamiltonian/inertial Monte Carlo
- Overregularization
- Distinctions between interpolation and extrapolation