AISEC: AI Secure and Explainable by Construction
Lead Research Organisation:
University of Edinburgh
Department Name: Sch of Informatics
Abstract
Abstracts are not currently available in GtR for all funded research. This is normally because the abstract was not required at the time of proposal submission, but may be because it included sensitive information such as personal details.
Organisations
Publications
Bogani R
(2022)
Garbage in, toxic data out: a proposal for ethical artificial intelligence sustainability impact statements.
in AI and ethics
Clausen H
(2021)
CBAM: A Contextual Model for Network Anomaly Detection
in Computers
Clausen H
(2021)
Examining traffic microstructures to improve model development
| Description | Since the start of AISEC, we have observed a flurry of activities by legislators in the UK and internationally to respond to the risks created by AI in general and autonomous driving and chatbots in particular. However, these legislative initiatives are frequently siloed and uncoordinated between different government departments, sub-national, national and international legislators. This can lead to regulatory gaps and normative conflicts between regulatory regimes, something we flagged up and evidence in detail our response to the UK government's consultation on future security challenges for AVs where we emphasised the need to coordinate in particular the post-Brexist data protection law with security-centric regulation. At the same time, third generation regulatory frameworks such as the proposed EU AI Act, the UK Regulatory white paper or the US Algorithmic Bill of Rights leverage "design solutions" and the role of programers and developers as new regulatory paradigm, creating new challenges for developers and regulatory burdens for industry. For formal verification, this also opens up new opportunities to assist industry in meeting this burden. Rather than representing road traffic law within the car- a strategy pursued by a number of similar projects - we demonstrated that it is possible to give a formal representation of the legal duties of AI developers. This allows smart developing environments with latent legal knowledge to not only ensure legal compliance in the planning and design stage, but also create legally sound documentation of the design decisions taken. The legal status of these "verifiably legal design processes" remains however controversial, and, again mirroring our recommendation to the UK government, may require a more radical rethink of the role of computer code in the legislative drafting process. |
| First Year Of Impact | 2022 |
| Sector | Government, Democracy and Justice |
| Impact Types | Policy & public services |
| Description | Independent Advisory Group on Emerging Technologies report |
| Geographic Reach | Local/Municipal/Regional |
| Policy Influence Type | Contribution to a national consultation/review |
| URL | https://www.sipr.ac.uk/wp-content/uploads/2023/02/Independent-advisory-group-on-emerging-technologie... |
| Description | Legislating for AVs in Dubai |
| Geographic Reach | Asia |
| Policy Influence Type | Participation in a guidance/advisory committee |
| Description | Submission to the UK Governemnt co nsultation on Self-driving vehicles: new safety ambitions |
| Geographic Reach | National |
| Policy Influence Type | Contribution to a national consultation/review |
| URL | https://www.gov.uk/government/consultations/self-driving-vehicles-new-safety-ambition |
| Title | MetaCP: Cryptographic Protocol Design Tool for Formal Verification |
| Description | MetaCP is a tool aiding the cryptographer throughout the process of designing and modelling a communication protocol suitable for formal verification. The crucial innovative aspect of the tool is its data-centric approach, where protocol specification is stored in a structured way rather than in natural languages. Starting with a modern graphical interface, the cryptographer can design a protocol with MetaCP and simultaneously export it into multiple target languages. |
| Type Of Material | Improvements to research infrastructure |
| Year Produced | 2020 |
| Provided To Others? | Yes |
| Impact | MetaCP is a first-of-its-kind tool aiding the protocol expert through the process of design and formal specification. It is based on a data centric approach where the protocol specification is described as a data structure containing all the information that must currently be interpreted from many pages written in English. It is a worthwhile exercise to attempt to standardize the methodology in which future (and current) protocols are designed. Formally verified models confer to protocols the reliability required to focus on the problems that raise at the implementation level. The methodology of MetaCP is a promising step towards this result. |
| URL | http://metacp.eu |
| Title | A Generic Methodology for the Statistically Uniform & Comparable Evaluation of Automated Trading Platform Components |
| Description | The primary objective of this research is to shed light on this field by providing a generic methodology that is investigation agnostic and interpretable to a human trader, thus reducing barriers to entry and increasing the reproducibility of experiments. |
| Type Of Material | Computer model/algorithm |
| Year Produced | 2022 |
| Provided To Others? | Yes |
| Impact | A publication |
| URL | https://zenodo.org/record/7011938 |
| Title | Machine Learning Classification of Price Extrema Based on Market Microstructure Features. A Case Study of S&P500 E-mini Futures. Reproducibility Package. |
| Description | The study introduces an automated trading system for S&P500 E-mini futures (ES) based on state-of-the-art machine learning. Concretely: we extract a set of scenarios from the tick market data to train the model and further use the predictions to model trading. We define the scenarios from the local extrema of the price action. Price extrema is a commonly traded pattern, however, to the best of our knowledge, there is no study presenting a pipeline for automated classification and profitability evaluation. Our study is filling this gap by presenting a broad evaluation of the approach showing the resulting mean annual Sharpe ratio of 6.32 for the period of 33 months. However, we do not take into account order execution queues, which of course affect the result in the live-trading setting. The obtained performance results give us confidence that this approach is worthwhile. Repository content * Original trained models are available in the `models` folder (unpack) * Classification dataset samples are available for each contract in the `datasets` folder (unpack) * Statistics for the model performance are available for your convenience in json format in the `fit_stats` folder (unpack) * Analysis of the feature performance is available in the `shaps` folder (unpack) To link these results to those of the publication please look at the `feature_names_association.json` file (unpack) to see which features correspond to those described in the paper· |
| Type Of Material | Computer model/algorithm |
| Year Produced | 2020 |
| Provided To Others? | Yes |
| Impact | A paper was published |
| URL | https://zenodo.org/record/4036851 |
| Description | Edinburgh Science Festival |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Public/other audiences |
| Results and Impact | National Robotarium Daytime Demo Exhibition: Autonomous Racing Car - We showcased the autonomous vehicle we built at the national science gallery in Edinburgh. The event boasts ove 100,000 annual visitors, with particular engagement with young kids and families. |
| Year(s) Of Engagement Activity | 2022 |
| Description | Interview with Direct Industry Magazine "Is the Dream of Self-Driving Cars Dead?" |
| Form Of Engagement Activity | A press release, press conference or response to a media enquiry/interview |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Industry/Business |
| Results and Impact | As an outcome of an article we wrote In collaboration with colleagues, I was contacted to interview for a magazine article on autonomous vehicles |
| Year(s) Of Engagement Activity | 2021 |
| URL | https://emag.directindustry.com/2021/10/26/is-the-dream-of-self-driving-cars-dead/ |
| Description | Invited talk at the First Annual LAST-JD-RIOE Conference of the Marie Sklodowska-Curie Action program "Rights in Internet of Everything" |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Postgraduate students |
| Results and Impact | The event was part of the Marie Sklodowska-Curie Action program "Rights in Internet of Everything", a EU project that sponsors a 3 year PhD program involving 15 European Universities It was attended by past and current students on the program, and also other PhD and Master students from participating institutions. I received several expressions of interest to come to the UK for a PhD in this field |
| Year(s) Of Engagement Activity | 2021 |
| URL | https://rioe.oeg.fi.upm.es/conference/#contact |
| Description | Magazine Article - The Conversation : "Perfecting Self Driving Cars - Can it be Done?" |
| Form Of Engagement Activity | A magazine, newsletter or online publication |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Media (as a channel to the public) |
| Results and Impact | As a general media piece, in collaboration with other authors from the project we wrote an article relating to our work with autonomous vehicles, over 25 thousand people read it, and it led to follow-up discussions with other media outlets and industry. |
| Year(s) Of Engagement Activity | 2021 |
| URL | https://theconversation.com/perfecting-self-driving-cars-can-it-be-done-157483 |
| Description | Participating in ReMeP workshop as invited particpant on roundtable on "Computational Law" |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Industry/Business |
| Results and Impact | Participated in a panel discussion on computational law as part of ReMeP - research meets practice. The aim was to introduce governments and legislators to the potential of computational law, and industry on the business opportunities to support this new form of lawmaking. The panel was attended by 140 participants A follow up request from the Austrian government to explore the possibility of a computational version of the GDPR |
| Year(s) Of Engagement Activity | 2021 |
| URL | https://remep.live/ |
| Description | Seminar on legal ontologies given at ReMeP: Research Meets Practice |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Industry/Business |
| Results and Impact | ReMeP is an annual event to bring research in law and technology to practice. It is attended mainly by representatives from industry, legal practice and government Chaired panel and gave keynote talk on Legal Ontologies. 470 participants, lead to extended questions and discussions, and 2 follow up invitations Recordings of the talks are made available as training tools |
| Year(s) Of Engagement Activity | 2020 |
| URL | https://www.remep.net/materials-2020/ |