Home  »  AI-SDV  »  IC-SDV 2019  »  Programme  »  Monday 8 April 2019

Monday 8 April 2019

09:00 - 10:30

Chair: Philip Ditchfield, DEEP SEARCH 9, UK

09:00 - 09:05

Conference Opening

Neural Networks today

Semantic Search can be understood in a  broad spectrum  in computer science in general, and in the intellectual property in particular. Among the methods  available to us, it has become clear that neural networks, be they deep, convolutional, recurrent, or otherwise, are the instrument of choice today. This is because, under the current data and computational resources availability, they are an extremely flexible instrument shown, repeatedly and consistently, to provide better results in a variety of machine learning tasks.


Thus, the aim of the presentation is to focus from the broad spectrum of semantics presented last year, to a narrow, but still very rich spectrum of neural networks. Additionally, I will show the directions in which the field appears to be developing, namely Transfer Learning and Differential Programming. Additionally, the presentation would focus on the need for standardised benchmarking and introduce the World Patent Information Patent Test Collection.




AI meets IP: There is Nothing Artificial about it

Artificial intelligence is a global phenomenon, a technology that has arrived. No industry will be untouched by the changes and disruption these technologies bring. With the rapidly changing innovation landscape, patent offices are discussing the interplay between AI and patents. Corporate directors, CEOs, vice presidents, managers, team leaders, entrepreneurs, investors, coaches, and policy makers are anxiously racing to learn about AI: they all realize it is about to fundamentally change their businesses. Patent analysts will have to respond to this changing environment by being more global in their perspective and will need analytic skills to deal with growing amount of data. The presentation will focus on these aspects and will highlight recent developments in AI methods and the breadth of AI applications that are of importance to patent searchers, analysts, and decision-makers.  We will discuss some basics of AI and then zoom in on the neural networks based natural language processing methods and discuss their applications for patent corpus.




10:30 - 11:00

Exhibition and Networking Break

11:00 - 12:30

Chair: René Deplanque, Germany

Down-to-earth machine learning: What you always wanted your data scientists to solve but were too busy to ask

Applications of machine learning on NLP tasks today receive a lot of attention and have been shown to yield state of the art results on a wide range of tasks. We describe several cases where machine learning is deployed productively under the usual constaints of real-world projects: Real-world requirements, fast throughput, reasonably low requirements in terms of training corpus size and high quality results. What we observe is a general trend towards open source - also our components are open source. With the software being mostly freely available, among the key success criteria for many NLP projects today therefore is first and foremost the necessary expertise required to combine, tune and apply open source components.

The RISE of Machine Learning; Predicting Future of Process Development


Process research and development of active pharmaceutical ingredients (API’s) plays an important role in the pharmaceutical industry. In this phase, all aspects of an API’s synthetic route are tested and optimized.

We envision that artificial intelligence and machine learning techniques could help in the optimization process of chemical transformations. In this presentation, the latest advances in the prediction of reaction conditions using machine learning algorithms will be discussed.

12:30 - 14:00

Lunch, Exhibition and Networking

14:00 - 15:40

Chair: Harald Jenny (CENTREDOC, Switzerland)

The Economics of Artificial Intelligence and Machine Learning for Automatic Categorization and Semantic Enrichment

Decision points for when to implement automatic indexing or more intensive subject analysis

Machine learning and artificial intelligence approaches to automatic indexing and other aspects of content enrichment have tremendous potential, but there are significant barriers to successful implementations.  The economics of these systems are not now generally affordable, which will indefinitely delay widespread adoption.  Significant costs are involved in just the training and maintaining systems that chronically under perform and are fail to scale.  Cost and performance data will be characterized and presented.  Machine learning and artificial intelligence projects are not for the faint of heart, nor for those with small budgets.  Key cost elements are identified along with approaches to estimating costs based on actual and reported cases.

Are Ontologies relevant in a Machine Learning World?


The unescapable rise of machine learning (ML) and artificial intelligence (AI) challenges the role of existing text analytics techniques such as Named Entity Recognition and Natural Language Processing in extracting information from scientific text. Often these rely on underlying ontologies to provide the semantic foundation for more complex linguistic and statistical analysis. This paper investigates how ontologies and ontology-led text analysis fits with emerging ML/AI algorithms and the synergies brought by combining the two approaches. We highlight real-world use-cases from across the Pharmaceutical and Life Science sector where SciBite’s text analytics systems have been employed to create next-generation enterprise data infrastructure for many of the world’s leading companies.


15:40 - 16:10

Exhibition and Networking Break

16:10 - 17:40

Chair: Christoph Haxel (Dr. Haxel CEM, Germany)

Technology fields supported by machine learning: A case study

 A recurrent topic in patent landscaping, monitoring and analysis is the necessity to operate in a defined technology field. The definition of such a field is, however, not always straightforward due to the ever increasing data volume and complexity. In addition, a traditional, classification based approach is often insufficient to encompass the relevant data in a growingly interdisciplinary context.

A number of predefined, precision rather than recall oriented technology fields is being developed by ip-search and rendered available within the PatentSight business intelligence platform. Further specificity can be achieved, if desired, via the application of a machine learning system.

A case study is presented. 

Distributing AI to the Amazon Cloud

The sheer size and number of websites that DS9 processes for its customers has meant the need for a new approach – one that dynamically scales available hardware resources and bandwidth. Amazon EC2 cloud resources was the platform of choice – and enables us to roll out DS9 installations as needed and execute jobs faster in the cloud. This talk presents the solution and also highlights some of the challenges that had to be overcome during the implementation phase.