Oltramari, Alessandro (Bosch Research and Technology Center, Pittsburgh (CR/RTC3.1-NA)) | Szurley, Joseph (Bosch Research and Technology Center, Pittsburgh (CR/RTC3.1-NA)) | Das, Samarjit (Bosch Research and Technology Center, Pittsburgh (CR/RTC3.1-NA)) | Francis, Jonathan (Bosch Research and Technology Center, Pittsburgh (CR/RTC3.1-NA), Carnegie Mellon University) | Li, Juncheng (Bosch Research and Technology Center, Pittsburgh (CR/RTC3.1-NA), Carnegie Mellon University)
In this position paper we discuss the benefits of combining knowledge technologies and deep learning (DL) for audio analytics: knowledge can enable high-level reasoning, helping to scale up intelligent systems from sound recognition to event analysis. We will also argue that a knowledge-integrated DL framework is key to enable smart environments.
In our new book Longevity Industry 1.0 - Defining the Biggest and Most Complex Industry in Human History, Dmitry Kaminskiy and I distill the complex assembly of deep market intelligence and industry knowledge that we have developed over the past five years. We present the full-scope of the Global Longevity Industry showing exactly how we managed to define this complex and multidimensional industry and how we created a tangible framework for systematization and forecasting. This article includes the preface and the foreword from our book. Every cause of death is, in principle, preventable. I became aware of this in the 1990s and since then I have been following aging research very closely.
Representation learning methods that transform encoded data (e.g., diagnosis and drug codes) into continuous vector spaces (i.e., vector embeddings) are critical for the application of deep learning in healthcare. Initial work in this area explored the use of variants of the word2vec algorithm to learn embeddings for medical concepts from electronic health records or medical claims datasets. We propose learning embeddings for medical concepts by using graph-based representation learning methods on SNOMED-CT, a widely popular knowledge graph in the healthcare domain with numerous operational and research applications. Current work presents an empirical analysis of various embedding methods, including the evaluation of their performance on multiple tasks of biomedical relevance (node classification, link prediction, and patient state prediction). Our results show that concept embeddings derived from the SNOMED-CT knowledge graph significantly outperform state-of-the-art embeddings, showing 5-6x improvement in ``concept similarity" and 6-20\% improvement in patient diagnosis.