AI in Government and Public Sector Fall Symposium Call for Participation

Deadline August 5, 2020 – https://easychair.org/conferences/?conf=fss20

AI is becoming ubiquitous, being useful across societal, governmental, and public sector applications. However, AI in Government at the federal, state, and local levels, and related education and public heath institutions (hereafter referred to as Public Sector) faces its own unique challenges. AI systems in the public sector will be held to a high standard since they must operate in support of the public good. They will face increased scrutiny and stringent requirements for ethical operation, accountability, transparency, fairness, security, explainability, cost-effectiveness, policy & regulatory compliance, and operation without unintended consequences.

We will invite thoughtful contributions – papers, speakers, and panel proposals – that present novel technical approaches to meeting these requirements and lessons learned from current implementations. We hope to provide some coverage on the use of AI to respond to the COVID-19 and Fairness, either through paper presentations, panel discussions, or invited Keynotes.

Potential topic areas include (in no particular order):

1. Technical Papers that advance the state-of-the-art on applying AI in the Public Sector Innovative approaches to solving the problems of building applications that meet the challenges described above.

o Responsible, Safe, and Trustworthy – Responsible use of AI is one of the important drivers in the successful deployment of AI. How can systems be designed to avoid bias in data, algorithms, and/or use of the system? How to ensure transparency, comprehensibility, and trustworthiness? This topic includes areas such as Open Data and Accountability of AI systems, as well as Open Source Code and Open Models and Training Methods.

o Verification and Validation for Deep Learning– Validation of AI models (especially those based on deep learning and other statistical machine learning techniques) and keeping them validated as the domain or model evolves, is an active topic of research. However, the public sector has a unique challenge in that often the correctness of the classifications that a deep learning model implements is ultimately derived from regulation or some other complex text. How do we validate these models, when human interpretation is an essential part of the correctness criteria?

o Privacy – This topic encompasses privacy issues in AI-based models, as they relate to GDPR (General Data Protection Regulation) and other National and State regulations, compliance, enforcement, and penalty issues. How to balance privacy needs with the public safety in which data available can lead towards increased benefits for many?

o Robustness and Resiliency –AI systems for the public sector must be robust and resilient in the face of the type of external manipulation, deception, and systemic cyber threats that the public sector uniquely faces. While Open Code requirements allow the detection of back doors and other problems, how this will work with trained AI models? How can we harden AI- based models from model poisoning and other adversarial attacks that are designed to misdirect or bias results?

o Public Sector Interaction Paradigms – This topic encompasses insights about various paradigms for AI use in public sector operations, such as intelligence augmentation and human-machine collaborative approaches, levels of autonomy, methods for handling uncertainty / conflicting evidence and opinion, and the various types of public sector users (government employees, general public, elected officials).

o Leveraging AI Innovation in Open Source – There are hundreds of open source AI projects, focusing on AI sub-domains such as deep learning, machine learning, models, and many more. How can government entities leverage the abundance of open source AI projects and solutions in building their needed platforms and services so that trusted and robust open source projects can be deployed for public service?

o Operation and Adaptation to Multiple Domains – Many public sector agencies must operate across multiple domains. For example, FEMA responds to hurricanes, earthquakes, and now the Corona virus pandemic. How can AI systems be made to support or adapt to cross domain operations? What are the challenges and research directions?

AI in Government and Public Sector Fall Symposium Call for Participation

2.

Practice Papers that describe current uses of AI in the public sector – applications which are early adopters of AI, role of public/private partnerships in accelerating development and adoption, timely response to societal challenges such as the COVID-19 Pandemic response (Public health, medical, social, economic), social justice, and demonstrations of beta and in-production applications.

o Early areas for adoption of AI – What public sector problems exist where AI is playing a large/important role without deep new experimentation or solving R&D problems? How can socially important challenges such as responding to COVID-19, fighting terrorism, better serving vulnerable populations, combatting racism, and so forth, be re-conceptualized to leverage AI’s strengths?

o Role of Public-Private Partnerships–What is the role of public-private partnerships in researching, creating, and operating AI systems for the public sector? How do AI R&D institutes being created with academia and industry provide a conduit for early adoption and transition of AI technologies in government?

o Using AI to encourage public service innovation – What public sector areas are not immediately approachable by AI, but still pose an urgent need, and hence offer a sufficient financial or social-good reward to justify investment and experimentation by public administrators?

o Translating from .com to .gov – What best practices and approaches can be transferred from the .com experience to benefit .gov?

o Systematic Approach for the Use of AI in the Public Sector–This topic encompasses policies, methodologies, guides and elements in support of public sector use of AI. In deploying AI technologies to improve public sector operations, tensions exist between effectiveness and protecting ownership and control rights to information. What are these tensions and trade-offs, and how can they be addressed?

o Cultivating AI Literacy – How do we facilitate understanding and acceptance of AI in the public sector? How do we start the conversation between government and citizens?
o AI Engineering Best Practices – The increasing prevalence of machine learning in

automation exposes AI to real-world data, and raises concerns about data drift, data poisoning, adversarial AI, and more. The complexity of modern probabilistic models and data pipelines raises the cost of understanding a system well enough to fix it when it breaks. These diverse concerns urgently call for AI engineering guides to help ensure robustness and transparency in AI solutions, with the cost effectiveness that is demanded of the public sector. What new best practices or standards are needed for engineering AI?

o Incentivizing AI Engineering Best Practices – The ability of the public sector to leverage AI depends in part on the availability of AI implementations that attain the highest levels of transparency, in terms of the documentation, the modularity of implementation, and adherence to potential standards. How should the public sector incent appropriate discourse and resolution of these issues? Who should do this?

Submissions

The symposium will include presentations of accepted papers in both oral and panel discussion formats, together with invited speakers and demonstrations. Potential symposium participants are invited to submit either a full-length technical paper or a short position paper for discussion. Full-length papers must be no longer than eight (8) pages, including references and figures and are required those submitting Technical Papers as described above. Short submissions can be up to four (4) pages in length and can be used for Practice Papers as described above, work in progress, system demonstrations, or panel discussions.

Please submit via the AAAI EasyChair.org site choosing the AAAI/FSS-20Artificial Intelligence in Government and Public Sector track at https://easychair.org/conferences/?conf=fss20 Please submit by August 5. Contact Frank Stein (fstein@us.ibm.com) with any questions.

Organizing Committee

Frank Stein, IBM (Chair); Erik Blasch, USAF; Mihai Boicu, GMU; Lashon Booker, Mitre; Michael Garris, NIST; Mark Greaves, PNNL; Eric Heim, CMU-SEI; David Martinez, MIT-LL;Tien Pham, CCDC ARL; AlunPreece, Cardiff University; Peter Santhanam, IBM; Jim Spohrer, IBM;