After-Action Review for AI (AAR/AI)

Explainable AI is growing in importance as AI pervades modern society, but few have studied how explainable AI can directly support people trying to assess an AI agent. Without a rigorous process, people may approach assessment in ad hoc ways—leading to the possibility of wide variations in assessment of the same agent due only to variations in their processes. AAR, or After-Action Review, is a method some military organizations use to assess human agents, and it has been validated in many domains. Drawing upon this strategy, we derived an After-Action Review for AI (AAR/AI), to organize ways people assess reinforcement learning agents in a sequential decision-making environment. We then investigated what AAR/AI brought to human assessors in two qualitative studies. The first investigated AAR/AI to gather formative information, and the second built upon the results, and also varied the type of explanation (model-free vs. model-based) used in the AAR/AI process. Among the results were the following: (1) participants reporting that AAR/AI helped to organize their thoughts and think logically about the agent, (2) AAR/AI encouraged participants to reason about the agent from a wide range of perspectives , and (3) participants were able to leverage AAR/AI with the model-based explanations to falsify the agent’s predictions.

Creation and Operation of a COVID-19 Pooled Testing Collection Site Prior to a CTC Rotation

ABSTRACT Introduction This brief report describes the process, used by the 1st Infantry Division (1ID) and Irwin Army Community Hospital (IACH) at Fort Riley, Kansas, for conducting pooled testing collection of over 2,500 Soldiers prior to a large-scale exercise involving multiple units. Materials and Methods The authors captured after action review comments on the process and results of their pooled specimen collection site. Pooled specimen test results were reviewed and classified according to Aberdeen Proving Ground criteria to determine the percentage of successful and failed pooled specimens. Results 1ID and IACH performed pooled testing collection and shipment of 2,684 specimens divided into 298 pools over 6 flight manifests. Of the 298 pooled specimens, 4 (1.34%) were found to be inconclusive or invalid, and the other 294 (98.7%) had sufficient number of human cells to be certified as SARS-CoV-2 (COVID-19) positive or COVID-19 not detected. Conclusion Pooled testing collection is a complex process that may continue to be a requirement for mass screening of COVID-19 prior to military operations. While planning should be tailored to the specific mission and unit, key factors that the authors feel are required for pooled testing to be successful in any situation are standardized training and personnel continuity, quality assurance, administrative oversight by the unit, and collaboration and communication between all involved entities.

Annulus Perforate, Wash and Cement : Establishing Barrier Verification Process for Deepwater Subsea Wells Abandonment and Its Lessons Learnt

Wells plug & abandonment was carried out in a deepwater field (Field C) offshore West Africa. There were 15 deepwater subsea wells, in this field. Thirteen of the wells were completed with Open Water Vertical Xmas Tree (OXT) while remaining two were completed with Enhanced Vertical Xmas Tree (EVXT). All the wells were permanently abandoned with permanent barriers established in accordance to Norsok D-010, rev. 2013. This involved establishing well barriers which would both horizontally and vertically effective. The operator evaluated several methods in establishing well barriers for the permanent abandonment purpose of these 15 wells. The barriers placement for areas with good annulus cement were straightforward as an internal cement plug could be spotted as a permanent barrier. For this particular field, it is noted that all of the 15 wells have hydrocarbon bearing and permeable formations above the top of expected annulus lead cement which is designed basically to manage ECD during cementing job rather than giving good annular isolation. Due to this, it is important for the operator to employ the best technique that could establish a permanent cement barrier both inside the casing and in the annulus at the same time. Operator learned that Section Milling was the widely accepted method to establish barrier in cases where there is poor cement or the annulus cement is not present particularly in land and offshore wells. However, this technique poses severe operational risk especially in a deepwater operation mode. Upon review, it was determined that Perforate, Wash and Cement (PWC) technology has been proven and it would become a valuable technique in establishing well barrier which complies with abandonment requirement as per the industry guidelines. High risk activities for deepwater operations such as section milling could be avoided by employing this method. Nonetheless, the Perforate, Wash and Cement (PWC) technology has its own set of risks especially with its method of establishing the well barrier immediately upon setting it. This involves drilling and re-logging the well immediately after setting the cement in annulus to prove the barrier. Upon discussion with PWC contractor, it was agreed that the verification technique may create additional hazards as it could jeopardize the integrity of the installed barrier in the annulus. The additional requirement for cement verification process will also prolong the project duration, hence is an additional cost to the project too. Operator worked together with the Perforate Wash and Cement (PWC) Contractor and Well Examiner to determine the best verification method for PWC plugs with the objective to comply with regulation on cement verification while at the same time, resulting in most cost-efficient and time-efficient operation. The establishment of Element Acceptance Criteria as an alternative Barrier Verification Process for PWC technology is necessary to improve the reliability of the system as well as reducing associated risks. In total, 18 PWC operations were successfully carried out throughout the campaign. An extensive after action review was conducted at the completion of each job to capture the valuable lessons learnt. These lessons learnt are shared in this paper as well which could add value for future operations with similar settings. It is expected that this paper will serve as a reference in the establishment of Element Acceptance Criteria as well as its implementation during the P&A campaign. It is hoped that the lessons learnt shared in the paper could assist other Operators on planning for similar campaigns in the future.

Effect of after action review on safety culture and second victim experience and its implementation in an Irish hospital: A mixed methods study protocol

Background After Action Review is a form of facilitated team learning and review of events. The methodology originated in the United States Army and forms part of the Incident Management Framework in the Irish Health Services. After Action Review has been hypothesized to improve safety culture and the effect of patient safety events on staff (second victim experience) in health care settings. Yet little direct evidence exists to support this and its implementation has not been studied. Aim To investigate the effect of After Action Review on safety culture and second victim experience and to examine After Action Review implementation in a hospital setting. Methods A mixed methods study will be conducted at an Irish hospital. To assess the effect on safety culture and second victim experience, hospital staff will complete surveys before and twelve months after the introduction of After Action Review to the hospital (Hospital Survey on Safety Culture 2.0 and Second Victim Experience and Support Tool). Approximately one in twelve staff will be trained as After Action Review Facilitators using a simulation based training programme. Six months after the After Action Review training, focus groups will be conducted with a stratified random sample of the trained facilitators. These will explore enablers and barriers to implementation using the Theoretical Domains Framework. At twelve months, information will be collected from the trained facilitators and the hospital to establish the quality and resource implications of implementing After Action Review. Discussion The results of the study will directly inform local hospital decision-making and national and international approaches to incorporating After Action Review in hospitals and other healthcare settings.

Management of a large outbreak of COVID-19 at a British Army training centre: lessons for the future

IntroductionThe COVID-19 pandemic has posed major challenges for infection control within training centres, both civilian and military. Here we present a narrative review of an outbreak that occurred at the Royal Military Academy Sandhurst (RMAS) in January–March 2021, in the context of the circulating, highly transmissible SARS-CoV-2 variant B.1.1.7.MethodsTesting for SARS-CoV-2 was performed using a combination of reverse transcriptase PCR and Lateral Flow Devices (LFDs). Testing and isolation procedures were conducted in line with a pre-established symptom stratification system. Genomic sequencing was performed on 10 sample isolates.ResultsBy the end of the outbreak, 185 cases (153 Officer Cadets, 32 permanent staff) had contracted confirmed COVID-19. This represented 15% of the total RMAS population. This resulted in 0 deaths and 0 hospitalisations, but due to necessary isolation procedures did represent an estimated 12 959 person-days of lost training. 9 of 10 (90%) of sequenced isolates had a reportable lineage. All of those reported were found to be the Alpha lineage B.1.1.7.ConclusionsWe discuss the key lessons learnt from the after-action review by the Incident Management Team. These include the importance of multidisciplinary working, the utility of sync matrices to monitor outbreaks in real time, issues around Officer Cadets reporting symptoms, timing of high-risk training activities, infrastructure and use of LFDs. COVID-19 represents a vital learning opportunity to minimise the impact of potential future pandemics, which may produce considerably higher morbidity and mortality in military populations.

Використання інтерактивних тривимірних візуалізацій для вивчення бойового досвіду підрозділів в операції Об’єднаних сил

Збір, обробка й аналіз об`єктивної та достовірної інформації для візуалізації бойових епізодів в зоні операції Об’єднаних сил, здійснюється з метою професійного аналізу дій протидіючих сторін та виявлення недоліків (або позитивних моментів) в діях наших військ. Це також являє собою важливе науково-практичне завдання для подальшого розвитку воєнного мистецтва ведення бойових дій на сході країни. Із зазначеною метою в країнах НАТО використовується методика After action review – аналіз проведених дій (АПД). У зв’язку з намаганням керівництва країни реформувати Збройні Сили України до стандартів НАТО, а за політичним планами, вступити до цього Альянсу, впровадження методики AAR (АПД) в ЗС України доцільно розглядати як актуальне завдання. Одним із ефективних засобів АПД є інтерактивна тривимірна візуалізація, яка включає в себе моделі місцевості й розташування на ній підрозділів (екіпажів, бійців) та з високою точністю й деталізацією відтворює хід бойових дій, що досліджуються у просторі й часі на загальному тактичному фоні. Для наукового дослідження й аналізу бойових епізодів, що уже сталися, та прогнозування можливих дій сторін в тій чи іншій бойовій ситуації, пропонується використання автоматизованого способу вибору раціонального сценарію бойових дій військових формувань сторін з однорідними бойовими засобами (наприклад, механізовані, танкові екіпажі та підрозділи).