At Caspian, we invest heavily in both internal and external research to further support the development of our AI technology solutions that are underpinned by advanced data science.
This includes publishing papers, building relationships with research organisations, knowledge transfer partnerships and opportunities for internships and studentships.
Following is a selection of our previously published research, white papers and case studies.
Balancing consistency and quality of investigation performance with operational efficiency is an ongoing challenge for financial services organisations. This paper explores how decision graph technology captures the analysis and decisioning traits of expert risk analysts to automate and augment investigations and deliver QA fail rate improvements of up to 80%.
What if your financial crime experts, instead of just training investigators, could also train system technology itself to do what the best investigators do when presented with an investigation? Our whitepaper in collaboration with Nasdaq shares previously unreleased research into the performance of risk investigation experts, how they make decisions and the process behind capturing a gold standard decision graph of such investigations.
Selected to appear in the proceedings of the International Joint Conference of Neural Networks (IJCNN) 2020, part of the World Congress on Computational Intelligence (WCCI). In the paper we present a Bayesian framework to stress test machine learning solutions. The framework is generic and agnostic to the modelling methods used.
A global Tier 1 Bank deployed automated AML technology as part of a focus on improving effectiveness and efficiency in the operation of their transaction-monitoring program. Read our case study produced in collaboration with Nasdaq to understand how the solution performed and the outcomes delivered.
Wherever machine learning is used in AML, there will be a model, and that model will carry risks that must be managed. This paper provides a comprehensive list of risks and mitigations that should be considered when managing model risk management.
In collaboration with Durham University’s Dr Noura Al Moubayed and published in the journal PeerJ Computer Science, this research is part of Caspian initiatives focused on natural language processing. The proposed approach has achieved very competitive performance to the state of the art with significantly less memory or computational expense.
We shine a light on the monitoring of money laundering during the COVID-19 pandemic and highlight the key issues that investigation teams in tier-one banks need to consider.
Knowledge Transfer Partnership
Knowledge Transfer Partnerships (KTP) provide us with opportunities to work with world leading academic teams and access advanced research that feeds into the further development of our automated investigation technology.
Knowledge Transfer Partnership with Durham University
We currently have a partnership in place with Durham University, one of the world’s leading research led universities, ranked 78th in the QS world rankings. We are working with the university’s Department of Computer Science (which itself is ranked 5th in the UK’s Complete University Guide) and are supported by Dr Noura Al-Moubayed, the academic project lead and Assistant Professor at Durham University specialising in Machine Learning and Natural Language Processing.
The collaboration with Durham university provides the Caspian team with access to discipline specific expertise that supports our work to continuously evolve and remain at the forefront of innovation in AI driven automated Anti-Money Laundering technology. The real-life financial crime problems we solve have wide reaching consequences both in a social and economic sense. The project focuses on automating the natural language explanations that our machine learning solutions output to help investigation teams in tier one banks become more effective and efficient.
As part of the Durham University Intensive Industrial Innovation Programme, Caspian have a PhD studentship in place that is fully funded by the European Regional Development Fund and is supported in their research by the data science team at Caspian.
The programme is designed to encourage innovation by supporting businesses (particularly in science and engineering) with a dedicated PhD research student for three years, as well as providing access to senior academic researchers and university research facilities.
With a focus on recent advances in deep learning, causal inference and natural language processing, the PhD provides the Caspian team with a way to think strategically and differently about about research objectives whilst having a material impact on our next generation of explainable machine learning technology.
Academic internships and placements
We regularly create opportunities for university students to take up short term internships and full year placements at Caspian.
Working with leading universities, these programmes give students of computer science and mathematics, a real-world opportunity to supplement their learning through participating in specific projects and shadowing the day-to-day operations of the Caspian data science team.
This provides extensive opportunity to test new ideas through proof-of-concept projects with the students which investigate and validate the potential of new developments to be worked into technologies within Caspian technology.
Enquiries and further resources
To enquire about research collaboration or partnership opportunities with Caspian, please contact us.
Our Resources page is also great way to follow updates from the Caspian team and includes our blog, news and events covering subjects from Data Science and AI to Financial Crime and Regulation.