physicsgraph

Summary of list:

  * Typer
  * Rich
  * Dear PyGui
  * PrettyErrors - formatting and coloring error messages in the terminal
  * Diagrams
  * Hydra - build configurations in a composable manner, and override certain parts from the command line or config files.
  * OmegaConf
  * PyTorch Lightning
  * Hummingbird - compile your trained traditional ML models into tensor computations
  * HiPlot - high-dimensional data using parallel plots
  * Scalene
  * Norfair - real-time object tracking; the real reason for this post, as it is created by the company

Did you try the Docker-based approach after trying Python's virtualenv? Or did you go directly to Docker?

A few years ago I set a goal to minimize the number of applications installed bare-metal on my laptop. Everything is containerized, just as the blog describes. There is some initial overhead, and some extra work associated with maintaining the infrastructure, but overall I can report improved reproducibility (due to an isolated and static environment per app).

I use a bash alias for each common application. For example, "jup" launches a jupyter notebook. I have a container with the Python package "Black" which runs using a git hook to clean up my code prior to commits.

My favorite quote: "prevent fast implementation of global changes." This is how large organizations become slower compared to small "nimble" companies. Everyone fails, but the noticeability of failure incentivizes large organizations to be more careful.

The same applies to pandemic preparedness. There is no incentive for a hospital to over-provision beds that are empty most of the time.

Looking for examples of cultures and organizations that over-provision, the military might be an example. Why do we translate worry about physical attacks into action but not cyber attacks or environmental harm?

This paper is well written. I was able to extract the mathematics into the project I created to document the relations and steps of derivations [0].

[0] https://derivationmap.net/review_derivation/608598/?referrer...

Many of the commenters here have pointed out that the semantic web did not become popular because the incentives are not supportive of the investment. I agree, and this leads me to wonder if there are domains in which the cost of investing in semantic markup are relevant.

I see scientific publishing as a venue that could clearly benefit from application of a semantic web-like domain-specific approach. There are a variety of possibilities [0], [1], but again many are implemented behind proprietary paywalls. There's value in search and knowledge representation, but the prohibitive cost of implementation is challenging.

[0] https://en.wikipedia.org/wiki/Semantic_Scholar [1] https://en.wikipedia.org/wiki/Web_of_Science

I taught Physics lab as a graduate student. I was surprised at how little formal guidance was provided by the University or department on how to teach effectively, how to engage with students, etc.

My guess about the level of engagement you are observing from the people you're paying: you are giving attention (measured in money) when few other people are demonstrating interest in their specialty.

The value of content tagging (e.g., PhysML) and keyword tagging (e.g., ScienceWise) is apparent in aggregate, like for searching. That benefit is to consumers while the burden is currently on the content creators.

I don't know of any incentives within academia or grant processes that would motivate content authors to tag their content. With the exception of creators of the tagging systems. That means bulk analysis (whether using a grammar or natural language machine learning method) is key.

Citation graphs (which is mostly what's been done previously[0]) pale in comparison to text analysis. The possibility of enabling complex searches would be a big leap forward in science.

[0] https://physicsderivationgraph.blogspot.com/2020/05/literatu...

There have been efforts to tag keywords in arXiv [0] and to identify sections of articles [1]. A conference on Mathematical Knowledge Management [2] was held last week; some participants have already been analyzing ArXiv.

Hopefully integration with Kaggle expands the number of teams taking advantage of the knowledge in the corpus.

[0] http://sciencewise.info/ [1] https://github.com/OMdoc/OMDoc/wiki/PhysML [2] https://cicm-conference.org/2020/cicm.php

That problem of pen-and-paper calculations featuring unintended errors is what I try addressing in a project I work on [1]. My approach is to use Sympy (which has a lot of Physics support) to validate expressions entered by a human. Not quite the AI-focus of this thread, but still a machine augmenting the work of researchers. To your point about the complexity of the math, the Physics Derivation Graph is able to handle simple inference rules but there's nothing preventing more advanced use.

[1] https://derivationmap.net/

I agree that navigating a map of Physics at the very lowest level would not enlighten any student or researcher. My expectation in mapping atomic steps for a wide swath of the domain might enable insights not otherwise accessible.

The chunking of atomic steps is what enables leaps in understanding. The mapping process starts with understanding each step.

The article is sparse on what a detailed map for mathematics would look like and merely points out that some topics have related techniques for solving them.

I don't think a map for math techniques is feasible, but a map relating topics via mathematical steps is possible in Physics [1]. (Disclaimer: I'm the author of that map for Physics.) I think the reason that a map in Physics is feasible is Physicists do not use math techniques in the way mathematicians do, and the objectives are different.

https://derivationmap.net/