DEV Community: 10kvClockman

Avatars in Simulations-A bit scary tbh but amazing: My second AI Update

10kvClockman — Tue, 31 Dec 2024 07:49:29 +0000

Mixing Live Hydra Interaction with Virtual Hydra Simulations
Prof. Jonathan Crego M.B.E
Director of the Hydra Foundation May 2021

In June 2020 I wrote a paper called Hydra Presence – the evolution of the Hydra methodology.

In that paper I explained the development work that had taken place to answer the following question:
How do we enable save-life professionals to gain access to Hydra training that is timely and relevant without the requirement for participants to travel or be co-located?

A huge amount of creativity from Adam Crego and his technical team who had already developed the latest Hydra in the Cloud technologies, created a new approach whereby all the existing methodologies of Hydra could be deployed in a virtual system. We called this new approach Hydra Presence. Participants are able to partake in a Hydra exercise, without the requirement of being co-located in a Hydra suite. We ran the first Hydra Presence workshop in December 2020 with University of Portsmouth and delivered their Masters in Police Science course. It was well received by the students, who quickly beame immersed in the exercise, working in virtual in teams, using Hydra tools, in virtual syndicate rooms. The students, represented by virtual avatars had lively discussions, over-talking each other, and working through the management of a simulated sudden death incident, considering their forensic strategies and jointly recording their decisons. From time- to-time all the students met in a virtual plenary room, where guided by their facilitators, differences of approach between different teams, were debated and compared. All of this occurred inside the shared virtual Hydra Presence environment, the students were geographically distant from each other, at locations all over the UK.

A video describing this event can be found at https://vimeo.com/516245370
The intensity of developing these capabilities has consumed us for over 18 months, they are a revolution in simulation technologies, whilst maintaining all that we have learned in the creation of Hydra methods over the last 30 years. We are incredibly proud of these new approaches and the team led by Adam, are now pushing the boundaries even further, as this paper is being written.

However
The technologies have been difficult to design and implement, (this is new territory after all), they have and are pushing the boundaries of immersive learning. However, at the end of the day, as sophisticated and revolutionary as they are, they must be seen as required and essential enabling technologies, through which Hydra methodologies can made deliverable, remotely. A separate technical paper will be forthcoming describing these unique tools. Consequently, this paper will not centre on these technologies, other than when referring to them as enabling methods to provide new approaches to immersive simulated environments. This paper is about the philosophy of using and harnessing these new tools in the creation and delivery of high fidelity, immersive and compelling simulations. It is about describing the creation of new Hydra methods, built on the shoulders of 30 years of Hydra experience and the collective wisdom of over 80 Hydra centres worldwide. It will keep true to the mission of Hydra; the examination of decision making within the context of save-life incidents, and in particular, the development of multi-discipline teams. Whilst ensuring the dynamics and tensions of inter-agency distrust and the realities of the asymmetry of power and influence, play out within high-fidelity high-stake critical incident management.

Now it gets interesting.
The creation of these technologies has enabled us to expand on the existing methodologies of Hydra. New tools have been designed to enable a completely new approach to Augmented Virtual Simulations. It is to this that this paper now turns its attention.
In the post-Covid reality, challenging questions have been asked about the added value of travel and face to face meetings. Or where logistically, it is difficult or impossible to meet at a location. Microsoft Teams ® or Zoom ® calls have evolved as useful tools for teams to share situational awareness and work together. These tools have had a huge impact and in the new post Covid normalities of distant working they are becoming standard working practices; however, they are not tools for simulations (although many innovative simulations have been delivered through them), they are not sufficiently engaging or interactive to replicate the dynamics of team-based decision making within the uncertainties and complexities of dynamic save-life problems.

Hydra Presence provides this environment and is being piloted in universities, the UK National Crime Agency and the London Metropolitan Police Force. These pilots are generating new capabilities for Hydra, they are important and exciting. Emerging from this research is the recognition that a more hybrid approach would reap additional. By Hybrid I am suggesting a mix between live in-person Hydra events where participants work together in syndicate rooms face to face and Hydra Presence students who are working in virtual syndicate rooms.

We have already successfully built methods where remote syndicates engage in Hydra Exercises. British Transport Police have a London Hydra Centre with remote syndicate rooms in Liverpool and Glasgow. Using this approach, participants in the remote syndicate rooms engage in the exercise as if they were down the corridor in the London suite. They participate in plenary session through video linking. This approach is now embedded in BTP and is well tested. There are however constraints. The remote room can not engage directly with individual participants in the central Hydra suite. The remote syndicate members respond as a group through the video link.

Hydra Presence removes these obstacles as each participant logged into the environment is able to engage with their fellow syndicate members directly as a group and individually with participants in the plenary discussions.

Using AI Avatars

10kvClockman — Tue, 31 Dec 2024 07:41:11 +0000

I have been experimenting with AI generated avatars with a view to incorporating them in my Hydra exercises. The AI voices are not good and do not contain the inflections, emotion and flow of human speech. Consequently, I tried using real speech and I am very happy with the outcomes. Here is the before and after.

https://vimeo.com/932785062/0dae9dda5c?share=copy

This was a video used in MACIE our multi agency child protection exercise.

Here is the new video generated by AI but utilising the original voice

https://vimeo.com/932785255/0d60b273f1?share=copy

I am very pleased with the results. The two-day exercise was run last month at Avon and Somerset Police HQ with all the professionals present. As always we delivered the exercise at no charge. At the end of the exercise we ran a 10kv debrief and analysed it with AIQA

2024 - Ultimate guide to LLM analysis using NLP standalone

10kvClockman — Mon, 30 Dec 2024 10:38:07 +0000

Title: Automated Thematic Analysis and Action Plan Generation Using NLP
Abstract: This paper outlines a novel methodology employing natural language processing (NLP) techniques to analyse debriefing workshop datasets. The workflow involves generating themes from participant text, associating text segments with themes, and synthesising actionable insights. The process is designed to systematically transform raw qualitative data into structured outputs for decision-making. All code was written in #php

Introduction: Analysing qualitative data from debriefing workshops is critical for deriving actionable insights. Traditional manual coding is labour-intensive and prone to subjectivity. This paper presents an automated workflow using NLP to streamline thematic analysis, align comments with themes, and produce actionable plans. Our approach leverages AI capabilities to ensure consistent, scalable, and high-quality outcomes.
One foundational framework informing this workflow is the "10,000 Volts Debriefing" method, developed by Professor Jonathan Crego. This approach emphasises immersive simulations followed by structured debriefing to extract insights from participants (Crego, "The 10,000 Volts Method"). Detailed descriptions of this methodology can be found on the LinkedIn profile of Jonathan Crego and the Hydra Foundation website (Hydra Foundation, n.d.). Incorporating principles from this framework ensures that the NLP-based thematic analysis aligns with best practices in debriefing.
Additionally, the use of AIQA (Artificial Intelligence for Qualitative Analysis), a system also developed by Jonathan Crego, strengthens the analytical capabilities of this workflow (Crego, "The Use of AIQA"). AIQA integrates structured inquiry techniques with AI models to support a deep analysis of qualitative datasets. It enables a dynamic interpretation of textual data, fostering robust insights tailored to decision-making scenarios. AIQA’s ability to handle large-scale qualitative datasets and embed structured inquiry principles ensures relevance and accuracy in deriving actionable insights.
Jonathan Crego MBE, a leader in immersive simulation and debriefing methodologies, has been instrumental in the development of AIQA and 10,000 Volts Debriefing. As the founder of the Hydra Foundation, his work emphasises multi-agency collaboration and critical incident training. His contributions to qualitative analysis and decision-making frameworks continue to influence practices globally, particularly in public safety and crisis management contexts.

Methods:

Data Preparation The dataset comprises anonymised text inputs from participants of debriefing workshops. Preprocessing involves: • Tokenisation: Segmenting text into meaningful units. • Noise Removal: Eliminating irrelevant content (e.g., stopwords, duplicates). • Text Normalisation: Converting text to lowercase and handling linguistic variations (e.g., stemming, lemmatisation).
Theme Generation 2.1 Initial Theme Extraction An AI model trained for topic modelling (e.g., Latent Dirichlet Allocation; Blei et al., 2003) is applied to: • Identify recurring themes across the dataset. • Output a preliminary list of themes and associated keywords. 2.2 Theme Refinement The AI-generated themes are further processed by the LLM, which consolidates overlapping or redundant themes into unique, finalised themes. This refinement step ensures semantic accuracy and contextual relevance.
Text-to-Theme Matching 3.1 Match Score Calculation Each paragraph is compared against the refined themes using the LLM to calculate semantic similarity. The model generates embeddings internally and computes similarity scores, which are expressed as percentages. This step ensures high accuracy and contextual relevance without relying on pre-trained external models. 3.2 Filtering Matches Themes with match scores above an adjustable threshold (default: 80%) are retained. The threshold is iteratively tuned to balance specificity and generalisability. Each theme is associated with a manageable number of comments, ensuring actionable insights.
Action Plan Development For each theme:
Key points from associated comments are synthesised.
An action plan is created, encompassing: o Key Points: Summarised insights from comments. o Action Points: Specific steps to address the theme. o Impact: Expected outcomes of the action points. o Measurement Measures: Criteria to evaluate success.
Final Report Generation 5.1 Embedding for Contextualisation Themes and their associated comments are sent to an embedding-based AI to enrich contextual understanding and ensure cohesive narratives. 5.2 Report Writing A text-generation AI (e.g., GPT-family model) generates the final report, including: • Thematic analysis overview. • Individual theme descriptions. • Synthesised action plans and conclusions. ________________________________________ Results and Discussion: We tested the methodology on a sample dataset of debriefing workshop texts. The LLM achieved over 90% accuracy in matching text to themes (validated through manual cross-checking). The action plans derived from AI outputs were deemed actionable and contextually relevant by domain experts. Key challenges included fine-tuning thresholds and addressing nuanced comments that required additional manual intervention. The inclusion of principles from the "10,000 Volts Debriefing" approach and AIQA methodology enhanced the interpretation of thematic analysis, enabling the process to incorporate real-world decision-making scenarios and critical incident frameworks effectively. The AIQA system’s integration ensured that the structured inquiry frameworks were maintained throughout the analysis. ________________________________________ Conclusion: This workflow demonstrates the potential of NLP in automating thematic analysis and action plan generation. Future work will focus on enhancing model explainability and exploring real-time applications in workshop settings. ________________________________________ Acknowledgements: We acknowledge the contributions of workshop participants and the support of advanced AI tools in implementing this methodology. References:
Blei, D. M., Ng, A. Y., & Jordan, M. I. (2003). Latent Dirichlet Allocation. Journal of Machine Learning Research, 3(4-5), 993-1022.
Crego, J. (n.d.). The Use of AIQA in Qualitative Analysis. Retrieved from https://linkedin.com.
Crego, J. (n.d.). The 10,000 Volts Method in Critical Incident Debriefing. Retrieved from https://linkedin.com.
Devlin, J., Chang, M. W., Lee, K., & Toutanova, K. (2019). BERT: Pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805.
Hydra Foundation. (n.d.). The "10,000 Volts" debriefing method. Retrieved from https://hydrafoundation.org.
Kudo, T., & Richardson, J. (2018). SentencePiece: A simple and language independent subword tokenizer and detokenizer for neural text processing. arXiv preprint arXiv:1808.06226.
Pedregosa, F., Varoquaux, G., Gramfort, A., et al. (2011). Scikit-learn: Machine learning in Python. Journal of Machine Learning Research, 12, 2825-2830.
Reimers, N., & Gurevych, I. (2019). Sentence-BERT: Sentence embeddings using Siamese BERT-networks. arXiv preprint arXiv:1908.10084.
Vaswani, A., Shazeer, N., Parmar, N., et al. (2017). Attention is all you need. Advances in Neural Information Processing Systems, 30.
van der Maaten, L., & Hinton, G. (2008). Visualising data using t-SNE. Journal of Machine Learning Research, 9(Nov), 2579-2605.
Wolf, T., Debut, L., Sanh, V., et al. (2020). Transformers: State-of-the-art natural language processing. Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing: System Demonstrations, 38-45.