Spotlight 01/22: AI for Arms Control. How Artificial Intelligence Can Foster Verification and Support Arms Control | References

    by Niklas Schörnig | To the Publication

    1 This text draws heavily on a longer chapter (Schörnig 2022) covering the same topic in a forthcoming anthology edited by Thomas Reinhold and the author entitled “Armament, Arms Control and Artificial Intelligence. The Impact of Software, Machine Learning and Artificial Intelligence on Armaments and Arms Control.” One of the seminal texts on the subject is still the PRIF Report “Machine Learning-powered Artificial Intelligence in Arms Control” (2019) by my former student and colleague Niko Lück, who, unfortunately, has left academia.

    2 Jasani and Barnaby 1984

    3 Din 1987

    4https://www.spiegel.de/netzwelt/gadgets/alphago-besiegt-lee-sedol-mit-4-zu-1-a-1082388.html

    5 If a text is, for example, in several columns or has captions, classical OCR software (optical character recognition software that transforms the image of characters into processable text) has difficulties identifying the correct sequence of the identified text. Trained AI can identify the correct text flow, reducing the human workload. Bast and Korzen 2017

    6 Gastelum and Shead 2018: 42

    7 Gastelum and Shead 2018

    8 Sundaresan, Chandrashekar et al. 2017

    9 Russel, Vaidya et al. 2010: 32

    10 Altmann 2020: 240-241

    11Boulanin, Brockmann et al. 2020: 5.

    12  Verbruggen 2022, forthcoming.

    13 Copy editing: Matthew Harris

     

    Further reading

    Altmann, J. (2020). Advances ind Seismic and Acoustic Monitoring. Nuclear Non-proliferation and Arms Control Verification. I. Niemeyer, M. Dreicer and G. Stein. Cham, Springer: 231-248.

    H. Bast and C. Korzen, "A Benchmark and Evaluation for Text Extraction from PDF," 2017 ACM/IEEE Joint Conference on Digital Libraries (JCDL), 2017, pp. 1-10, doi: 10.1109/JCDL.2017.7991564.

    Boulanin, V., et al. (2020); Responsible Artificial Intelligence research and Innovation for International Peace and Security, Stockholm: https://www.sipri.org/sites/default/files/2020-11/sipri_report_responsible_artificial_intelligence_research_and_innovation_for_international_peace_and_security_2011.pdf,

    Din, A. M., Ed. (1987). Arms and Artificial Intelligence. Oxford, Oxford University Press.

    Gastelum, Z. N. and T. M. Shead (2018). "Inferring the Operational Status of Nuclear Facilities with Convolutional Neural Networks to Support International Safeguard Verification." Journal of Nuclear Material Management XLVI(3): 37-47.

    Jasani, B. and F. Barnaby (1984): Verification Technologies- The Case for Surveillance by Consent. Warwickshire, Berg Publishers.

    Keir, D. and A. Persbo (2020). History, Status and Challenges for Non-proliferation and Arms Control Verification. Nuclear Non-proliferation and Arms Control Verification. I. Niemeyer, M. Dreicer and G. Stein. Cham, Springer: 15-26.

    Lück, N. (2019): Machine Learning-powered Artificial Intelligence in Arms Control, PRIF Report 8/2019, Frankfurt/M.

    Russel, S., et al. (2010): Machine Learning for Comprehensive Nuclear-Test-Ban Treaty Monitoring. CTBTO Spectrum 14.

    Schörnig, Niklas (2022): AI for Arms Control. In: Reinhold, T./Schörnig, N. (Eds.): The impact of software, machine learning and artificial intelligence on armament and arms control”. Forthcoming.

    Sundaresan, L., et al. (2017). "Discriminating  Uranium  and  Copper  Mills  UsingSatellite Imagery." Remote Sensing Applications: Society and Environment 5: 27-35.

    Verbruggen Maaike (2022): No, not that Verification: The problems with Testing, Evaluation, Validation and Verification of Artificial Intelligence in Weapon Systems. In: Reinhold, T./Schörnig, N. (Hrsg.): The impact of software, machine learning and artificial intelligence on armament and arms control“. Forthcoming.