MLighter

MLighter

Contact

If you want to use MLighter, or you need more information about the product, please contact us in: hector.menendez (at) kcl.ac.uk

Testing Machine Learning

Machine learning testing market is relatively new and hence the market size is currently unclear. However, it is possible to correlate the ML testing market with that of software testing to have an estimate of...

Developers

Our team covers different aspects of machine learning and cybersecurity and have years of experience in both areas. Our team focuses on developing new features for MLigther aiming to improve your testing experience. Héctor D....

Testing Machine Learning

Machine learning testing market is relatively new and hence the market size is currently unclear. However, it is possible to correlate the ML testing market with that of software testing to have an estimate of the market share. Our market research highlighted the fact that software testing corresponds to 10% of the overall software development market.

Furthermore, the widespread use of ML applications implies that the demand for ML is on the rise; it is expected that testing machine learning will make up to 10% of the whole machine learning market (currently standing at around 8.4B$, therefore the testing market would reach up to 840M$).

‘Within this specific market, we identified 7 different segments that cover 97% of the market. Out of these 7 segments, our market validation has focused on four segments namely, healthcare (13%), retail (12% of the market), IT (22% of the market) and automobile (15% of the market).

The aim of MLighter is to help machine learning coders and QA testers to make their machine learning systems more resilient to potential adversarial conditions and identify bugs and vulnerabilities in their code.

NEWS: Get ready for the Demo!

MLighter has been funded by Innovate UK under the CyberASAP program. The goal of this program is to transform a research idea into a product... Read More "NEWS: Get ready for the Demo!"

How MLighter was born?

Everything started with a paper on adversarial machine learning. During the period that Hector was working at UCL, he learnt about different applications of machine... Read More "How MLighter was born?"

Why do we collect ML Bugs?

In several cases, when we are working with programs, we identified bugs. These bugs are related to the functionality, the performance, or the security of... Read More "Why do we collect ML Bugs?"

Testing ML: Is that a thing?

Machine learning is changing the world. The development of this technology is leading us to the 4th industrial revolution where artificial intelligence is assisting humans... Read More "Testing ML: Is that a thing?"

Features

There are multiple features that MLighter provide that can help you on your machine learning testing process. We have organise some of these features depending on the profile that you have, mainly forcused on ML Developers and QA Testers:

For ML Developers:

  • Familiar coding interface
  • Testing code and models
  • Vulnerability hunter
  • Performance evaluation
  • Historical vulnerabilities

For QA Testers:

  • Easy-to-use-interface
  • Finding model blind-spots
  • Adversarial scenarios
  • Database of attacks
  • Explanatory Reports

For Everyone:

  • Reliable ML models
  • Improve ML Security
  • Understandable ML
  • ML Testing Community
  • Cloud Integration

Developers

Our team covers different aspects of machine learning and cybersecurity and have years of experience in both areas. Our team focuses on developing new features for MLigther aiming to improve your testing experience.

Héctor D. Menéndez

Hector D. Menendez is currently a lecturer in Computer Science at King’s College London. He is a computer scientist (BSc, MSc and PhD) and a mathematician (BSc and MSc). He started working in machine learning during his PhD but, during his postdoc at University College London (UCL) with Dr David Clark, he started exploring different areas of “Comprology”, mainly security, malware, diversity, and testing.

He is a researcher, developer and business manager. He is currently collaborating with multiple researchers all around the world in three research topics: malware, software testing and machine learning; as a developer, he is creating the MLigther system for holistically testing machine learning; and as a business manager, he is leading Endless Science, a new publisher that will reinvent open scientific publications.

Karine Even-Mendoza

Karine Even-Mendoza is currently a Lecturer in Systems and Programming Languages (CS) at King’s College London with a PhD in Computer Science from King’s College London. Before joining King’s as a lecturer, she was a Research Associate in the Department of Computing at Imperial College London, where she worked in the Software Reliability Group (SRG) and Multicore Programming Group on compiler testing and software testing in general with Prof. Cristian Cadar and Prof. Alastair F. Donaldson. She completed her PhD at King’s College London, studied there, and worked at the Software Systems (SSY) group for four years. Her PhD research on software model checking and abstraction refinement was done under the supervision of Dr Hana Chockler (first) and Prof. Luca Viganò (second) in collaboration with the USI Formal Verification and Security group. Her PhD thesis was on Efficient SMT-based Verification of Software Programs in the field of model checking, SMT solving, and incremental verification for software. Prior to her PhD studies, Karine worked in several local and international software companies as a research scientist and software developer.

David Kelly

David Kelly is currently a Post Doctoral Research Associate at King’s College London. He holds a PhD and MSc in Computer Science from University College London. His PhD, on the merger of information theory with security type systems and testing was done under the (joint first) supervision of Prof Earl T. Barr and Dr David Clark.
His research is at the intersection of formal methods and probabilistic reasoning, with a special interest in applications of information theory to engineering practice and security. At King’s he is working on statistical methods for explainable AI. He was not always a computer scientist: he holds a BA in Music from Trinity College, Dublin. Prior to his PhD, he worked for many years as a stone carver and lettering artist, having worked in St Paul’s Cathedral, London, and the National Memorial Arboretum (UK) among many other locations.

Contact

If you want to use MLighter, or you need more information about the product, please contact us in:

hector.menendez (at) kcl.ac.uk