Hidden Technical Debt In Machine Learning Systems

Hidden Technical Debt In Machine Learning Systems. Empirical analysis of hidden technical debt patterns in machine learning software. In a recent paper¹, a team of google researchers discuss the technical debt hiding in machine learning (ml) systems.

The hidden technical debt in machine learning systems (adapted from [6 from www.researchgate.net

Sculley and his colleagues at google came up with “hidden technical debt” (htd) framework [], to address maintainability issues of ml software.definition of the htd patterns that are the focus of this paper can be found in our online repository. Machine learning offers a fantastically powerful toolkit for building useful complexprediction systems quickly. Here we examine several ways that the resulting erosion of boundaries may significantly increase technical debt in ml systems.

Source: www.researchgate.net

This is an updated version of the previous article, which i reviewed here: Machine learning offers a fantastically powerful toolkit for building useful complexprediction systems quickly.

Source: www.pinterest.co.uk

Ml allows us to build useful complex prediction systems quickly, but this does not come for free. Due to the complexity of the models at hand, different problems can often emerge and lead to technical debts in machine learning systems.

Hidden Technical Debt In Machine Learning Systems 1 Offers A Very Interesting High Level Overview Of The Numerous Extra Layers Of Technical Debt 2 Which Exist In Machine Learning Enabled Systems.

Empirical analysis of hidden technical debt patterns in machine learning software. The problem with the technical debt though is the same as with the financial debt — when the time comes to pay the debt we give back more than we took at the beginning. Preliminary results indicate that emergence of significant amount of htd patterns can occur during prototyping phase, however, generalizability of the results require analyses of further ml systems from various domains.

Sculley And His Colleagues At Google Came Up With “Hidden Technical Debt” (Htd) Framework [], To Address Maintainability Issues Of Ml Software.definition Of The Htd Patterns That Are The Focus Of This Paper Can Be Found In Our Online Repository.

Technical debt is an instrument which is justified when we need to meet some release deadlines or unblock a colleague. This paper identifies the technical debts that are special to ml systems and provides some common mitigation strategies. The paper, hidden technical debt in machine learning systems, talks about technical debt and other ml specific debts that are hard to detect or hidden.

The Boundary Erosion Caused By Complex Model.

Here we examine several ways that the resulting erosion of boundaries may significantly increase technical debt in ml systems. That is because the technical debt has a compound effect. Machine learning systems often mix signals, entangling them and making isolation of improvements impossible.

This Debt May Be Difficult To Detect Because It Exists At The System Level Rather Than The Code Level.

Ml systems have a special capacity for incurring technical debt because they have all of the maintenance problems of traditional code plus. Many of these now begin to face common challenges that have only started being addressed. Hidden technical debt in machine learning systems is the title of a great clear straight to the point paper published by google.

Encapsulating Objects Enables Easier Code Maintenance.

Hidden debt is dangerous because it compounds silently. Ml systems can produce a number of issues previously not. This paper argues it is dangerous to think ofthese quick wins as coming for free.