New Analytics Entity (AI Industry Analytics)
EstablishedTo Conduct First-Ever Profiling
of Global AI Industry Ecosystem
New analytics subsidiary of Deep Knowledge Group (DKG) has developed first-ever methodology, framework and AI algorithm able to profile full AI Industry Ecosystem
Tuesday, May 2, 2023, UK: The largest-ever profiling to date of the global AI Industry is currently being conducted by a new agency dedicated to conducting sophisticated comparative analytics, forecasting and benchmarking of the global AI industrial, policy and technological ecosystem, AI Industry Analytics (AIIA), in what amounts to the first-ever profiling of the global AI Industry produced by an AI itself.
The agency has finished developing the core framework and methodology being used to profile and map the Global AI Industry Ecosystem, which is itself a first-of-its-kind achievement. AIIA has also at this stage finished development and validation of the proprietary suite of AI algorithms and systems that are being used to conduct their analyses. Having now applied these algorithms to the full-scale profiling of the global AI ecosystem, AIIA is now in the process of refining the big data data visualisation techniques that will be used to illustrate such an enormous amount of data in a way that is understandable and digestible for humans in a single display.
AIIA was launched as a spin-out of Deep Knowledge Analytics, a DeepTech-focused analytics agency with a long history of AI industry analytics (including first-of-their-kind AI industry mapping of the UK, Switzerland, and several other regions), and Deep Knowledge Group, a data-driven consortium of commercial and non-profit organizations known for their long-term technological forecasting capabilities, their sophisticated AI and Data Science Department, and their very broad range of activities in the realm of DeepTech and Frontier Technologies across scientific research, investment, entrepreneurship, analytics, consulting, media, philanthropy and more.
The challenge of developing an AI system capable of semi-autonomous data aggregation, parsing, structuring and analysis, and the extraction of actionable insights able to be used for comparative analyses, market trends identification, industry benchmarking and forecasting of the Global AI Industry Ecosystem is substantial. It required the development of a reinforcement learning-based NLP explicitly designed to extract key relations from massive volumes of unstructured text data, capable of both searching, extracting, analysing and structuring large amounts of AI industry data, and extracting key insights from that data, on the basis of text-prompts for sufficient task-universality.
“Our initial validation results indicate that the proprietary model developed by our Data Science team is better at accurate autonomous data aggregation and analysis on global AI industry activity than other models, such as ChatGPT4. In test trials we asked the model to retrieve information about fundings attracted by AI companies mentioned in the news articles, and found that our model has higher accuracy at this task than GPT4 - in particular, achieving an accuracy of 0.7843 compared to 0.6413 for ChatGPT4. The task objective was to find the target company, round type, date of round, money raised, lead investors, and other investors. Our data scientists benchmarked AIIA’s model against GPT4 and GPT3.5-turbo via API on our test data, which any model given this task didn't see. The final score is the average of accuracies for 200 examples. If a model correctly predicts all 6 objectives, the accuracy is 1. Still, for example, if the model does not find a complete list of other investors, it is considered a false prediction and the performance is 5/6.” commented Dominic Jennings, Head of Strategic Partnerships for Deep Knowledge Group / DeepTech and Longevity Industry Financial Advisors.
With roughly 1000 different types of relations built into the model, the task of extracting relevant information and correctly omitting non-relevant information was paramount. A key innovation of AIIA’s algorithm is in its capacity to accurately detect specific relations which are very sparse in a given volume of text and financial data. With such large volumes of data and such high degrees of sparsity in the information being targeted, small inaccuracies can quickly pile up into large numbers of false positives, fatally preventing relevant analysis.
“We found that for the specific use-case of AI Industry Ecosystem data aggregation and knowledge extraction, our data scientists were able to build a model which is ~10x smaller than common LLMs like ChatGPT, which has comparable and in some cases superior results to those well-known in terms of accuracy and performance, while still retaining comparable interpretability.” said Dmitry Kaminskiy, General Partner of Deep Knowledge Group.
Having finalized the methodology and frameworks for full AI industry ecosystem profiling, developed and validated the proprietary AI algorithms and models for automated data aggregation and knowledge extraction, and completed the practical execution of that industry ecosystem-wide mapping, the agency is now working on the final refinement of the Big Data visualization techniques that it will use to convey those insights to the general public and key AI industry decision-makers and stakeholders - a laborious process, given the unprecedented volumes of data they are working with.
“The AIIA data science team has done a remarkable job in profiling the full extent of the global AI ecosystem from the standpoint of investments, finance and industry, but this is only the beginning. Several similar projects mapping the AI policy, governance, regulation and ethics space are also in the works, and will be released in due course. Overall I feel that this project has been a long time coming for Deep Knowledge Group, which in 2014 (a year before I joined them as a Longevity Policy analyst) made international headlines for its decision to appoint an AI algorithm to its Board of Directors. Given that the AI policy, governance and ethics landscape holds more promise and potential peril for global society than for any other technological domain, I am delighted to see Deep Knowledge Group continue to prioritize (and in some sense return to) this all-important domain at the height of its AI analytics and Data Science capabilities” said Ian Inkster, Head of Policy Analytics for Aging Analytics Agency and Deep Knowledge Group.
About Artificial Intelligence Industry Analytics
AI Industry Analytics (AIIA) is a spinout of Deep Knowledge Analytics and Deep Knowledge Group created in 2023 in response to various explosive new developments in AI. AIIA combines the power of market analytics and investment data with advanced algorithms and machine learning techniques to provide unparalleled insights and solutions in real-time in order to provide valuable information to its clients on the latest trends and developments in the AI industry.
About Deep Knowledge Group
Deep Knowledge Group is a leading consortium of commercial and non-profit organisations active on multiple fronts in the realm of DeepTech and Frontier Technologies (AI, Longevity, FinTech, GovTech, InvestTech), ranging from scientific research to investment, entrepreneurship, analytics, media, philanthropy and more. As a mathematical corporation focused on building the bridge to the 5th Industrial Revolution, Deep Knowledge Group is decisively committed to DeepTech for Social Good, Techno-Philanthropy and DeepTech and Longevity Industry Financial Commoditization (establishing the core investment and financial industry infrastructure needed for the emergence of DeepTech and Longevity as fundamentally new asset classes). The consortium’s unique approach lies in combining ecosystemic financial innovations with sophisticated AI and Data Science technology development for DeepTech and Longevity Industry analysis, benchmarking and forecasting.