Motivation

How many design patterns? I guess there’s limited number for that? Once you DS “How many different categories of ML question?” The number is very shocking. 24 or More.

Another issue which talked from other posts is overdesign.

1. Can I find out 3 different design patterns are more useful in ML application?
2. When should we start to think of design patterns?
3. I can apply the the pesudo project for the exist desgin pattern.

Some Patterns:

• Layer
• Pipe/filter
• Event bus
• Model view control
• client- server
• broker

Design patterns for Machine Learning Applications

This essay discuesses few different common SAs.

• Model-View-Controller
• pipes and filterd model

and following 2 real problems:

• cancer cell detection
• intrusion detection

Question behind paper

What’s the difference between filterd model and layered model?

Model	PIPE/FILTERED	LAYERD
self-contained	Lower	Higher, only allow to use few common dependence
Enviroments	It might need to work on same languages	Should be work alone. Well-defined interface

When should be aware of the design pattern?

TBH, this paper dosn’t anwser my question at all. Can’t tell what’s the unique requirements difference between ML APPs/ DL APPs/ General APPs

Scikit-Learn Design Principles

This post doesn’t relate to the design patterns though. It enlight my mind about the advanced coding skill.

importance of programming to an interface, not an implementation.

Sciket-learn has very strong user example and default case. In most of first-trial, we can leave a plenty of configurations empties and get some resaonable(inperfect) result.

Good design must start from easy example and default settings.

Abstract factory groups object factories that have a common theme.

Estimator, Predictor, and Transformer. Importantly and crucially these interfaces are complimentary

Extensibility Through Duck-Typing

A programming style which does not look at an object’s type to determine if it has the right interface; instead, the method or attribute is simply called or used (“If it looks like a duck and quacks like a duck, it must be a duck.”) By emphasizing interfaces rather than specific types, well-designed code improves its flexibility by allowing polymorphic substitution. Duck-typing avoids tests using type() or isinstance(). (Note, however, that duck-typing can be complemented with abstract base classes.) Instead, it typically employs hasattr() tests or EAFP programming.

• Previous
  Designing Data-Intensive Applications Note
• Next
  Portfolio