Monday 4. October 2021

In the coming years, an immense number of people will retire from the workforce as the so-called "baby boomers" reach retirement age. This poses a major challenge to science and industry, because this is the generation that is responsible for much of the scientific and technical growth of the last three decades. Among other things, they have paved the way for modern genome-based medicine, such as genome sequencing, gene synthesis and the copying and amplification of genetic material. They revolutionized diagnostics, enabled genome editing with CRISPR/CAS and after all: they were the ones who started all that “digitalization” thing. But what happens to the knowledge and expertise of these professionals when they get retired? The risk of a collective loss of internal (organizational) memory is substantial. As organizations face the potential loss of precious knowledge, the question arises: how can this knowledge be captured and made available for future generations? How can we make sure that this knowledge is made reusable? Somehow, the expert "thinking" must be “computerized” and turned into a format that allows to make it available to others. The good news is: We have a solution for that. We use an easy-to-use authoring tool as a knowledge-digitization-guide and let professionals play “ping pong” without forcing them to deal with basic principles of formal knowledge representations. “Ping pong” is a metaphor for an innovative approach to "playfully" bring knowledge from the minds of experts into a reusable form (a graph) that can then be used by network algorithms and thus be considered “in the computer". Our integrated intelligence approach creates a synergistic collaboration between human beings and modern AI techniques. Initial starting point is a mind map, a well-established and widely used, simple tool for the capturing of ideas and concepts. The professional makes a first sketch of the knowledge representation that will be the subject of the “ping pong” match. This first overview of topics, ideas and concepts is the "ping", a request to the machine for supporting, context expanding answers (“pong”) to extend the graph step by step by new aspects. In this dynamic, iterative process, the concept overview becomes a highly enriched knowledge graph, which can then be added to a constantly growing Enterprise Science & Technology Graph.

AILANI is a novel and unique semantic search enterprise solution for fast, easy and comprehensive knowledge discovery. It combines semantic modelling, ontologies, linguistics and artificial intelligence (AI) algorithms in a self-refining system that delivers results based on inter-related meaning of facts. AILANI not just allows for phrase searches as well as structured queries, it offers its users a unique hybrid natural language question answering system combining machine learning algorithms with semantic network-based "prior knowledge" inference. It integrates seamlessly with existing infrastructure and helps leverage knowledge buried both in decade old data as well as data derived from news feeds and clinical trials providing real-time semantic analysis of breaking news. For the pharmaceutical industry it is critical to stay up to date with the latest clinical trials news for decision-making in drug development. Integration of the relevant data and using ontology-based refiners enables fast and efficient retrieval of information about the clinical competitive landscape.

AI is not something futuristic any more, but has become an integrated part of our day to day life in IP. Earlier on, AI applications were isolated elements of the searcher’s activities; today, they are fully integrated in the workflow of technology monitoring, allowing to undertake the step from Artificial to Augmented Intelligence.

The global data sphere, consisting of machine data and human data, is growing exponentially reaching the order of zettabytes. In comparison, the processing power of computers has been stagnating for many years. Artificial Intelligence – a newer variant of Machine Learning – bypasses the need to understand a system when modelling it; however, this convenience comes with extremely high energy consumption. 

The complexity of language makes statistical Natural Language Understanding (NLU) models particularly energy hungry. Since most of the zettabyte data sphere consists of human data, such as texts or social networks, we face four major obstacles:

1.            Findability of Information – when truth is hard to find, fake news rule

2.            Von Neumann Gap – when processors cannot process faster, then we need more of them (energy)

3.            Stuck in the Average – when statistical models generate a bias toward the majority, innovation has a hard time

4.            Privacy – if user profiles are created “passively” on the server side instead of “actively” on the client side, we lose control

The current approach to overcoming these limitations is to use larger and larger data sets on more and more processing nodes for training. AI algorithms should be optimized for efficiency rather than precision. In this case, statistical modelling should be disqualified as a brute force approach for language applications. When replacing statistical modelling and arithmetic, set theory and geometry seem to be a much better choice as it allows the direct processing of words instead of their occurrence counts, which is exactly what the human brain does with language – using only 7 Watts!

New technologies like CRISPR-CAS or mRNA based vaccines require CI teams to constantly screen the market for opportunities and threats. The need for more focused intelligence that is targeted to the company’s strategic plans and faster access to critical information for decision makers has been increasing dramatically: More opportunities can be turned into a competitive advantage and management can avert potential threats before they become problematic. Monitoring early technology development, finding licensing opportunities or acquisition targets as well as quick access to broad clinical trial information and close surveillance of the competition are key disciplines of CI teams in R&D.

Structured vocabularies, thesauri and lexicons are key ingredients for many information management tasks. Creating them however often requires a significant amount of work. Maintaining and extending them often means that the respective manual tasks need to be done on a regular basis in order to prevent the resources from becoming outdated, irrelevant and incomplete. AI has much support to offer for this task. And by wrapping the respective approaches into applications that can be operated by terminologists and domain experts who don't need to be programmers or data scientists themselves, the benefits can be made available to a wide range of users.