Defending the Thesis — Geometric Regularization from Overparameterization The preprintmedium.com 1) Response to reviewer 6bEn Originality We thank you for noting that the conjecture is fairly novel and proposes a new link between an established phenomenon in geometry and an empirical phenomenon in deep learning. …

Avoid an international incident — The increasing sophistication of automated language translation via natural language models may carry a hidden risk. Just as people began to drive faster after we installed seatbelts, the recent improved performance of algorithmic translation services may carry a risk of increased reliance on voracity of interpretations without human validation. By…

What to do when nothing can be done — The Calm The evening breeze feels slightly cooler on your shoulder today, yet a palatable kind of stillness lingers on the horizon. You almost have an inkling that the birds have begun to subtly disperse, the dogs no longer want to play outdoors. The falling sunset with a surrounding orange glow more…

The preprint — Abstract The volume of the distribution of weight sets associated with a loss value may be the source of implicit regularization from overparameterization due to the phenomenon of contracting volume with increasing dimensions for geometric figures demonstrated by hyperspheres. We introduce the geometric regularization conjecture and extract to an explanation for…

Make yourself explicit. — As this essay collection progresses beyond the Automunge project, am trying to avoid making it some kind of vanity project. The goal for some time now has been to contribute to field of computer science, to add value through whatever channels become available. Part of the risk of dedicated focus…

Every life has a story — An aphorism is a condensed form of writing in which one can find the theme of a novel, the lesson of an age, or a pithy diversion. No aphorism is complete without context, but one may feel free to bring their own. The cheapest way to sedate someone is to…

Ordered out of order — Machine Learning — the algorithmic derivation of a modeled data generating function Artificial Intelligence — a superset that includes machine learning and other conventions of modeling knowledge Supervised Learning — a type of machine learning that relies on the convention of a training phase comparing training data and corresponding labels…

Data types matter — One Looking back on the ICML conference, one of the eye opening takeaways to this author was actually associated with no more than a brief comment by a presenter. In Weinan E’s keynote Towards a Mathematical Theory of Machine Learning [1], he shared a slide commenting on the global minima of…

Stable version release 8.30 — Looking back, the Automunge development process I think was somewhat unique in the field for the sheer pace of iterative developments throughout the multi-year span of focus. It is not unsurprising that in this context the interest of a user base was not exactly what I would consider as compounding…