Archive

Within the context of text-as-data paradigm, new type of analysis and metrics are today available to the political science research, allowing new and different scientific approaches. Political science is an area particularly benefited with this new paradigm. Traditionally, all national parliaments have produced comprehensive data for the whole legislative process, that includes a faithful transcription of all parliamentary debates. These transcriptions constitute a rich database,...

Single-cell sequencing technology holds the promise of unravelling cell heterogeneities hidden in ubiquitous bulk-level analyses. However, limitations of current experimental methods pose new obstacles that prevent accurate conclusions from being drawn. To overcome this, researchers have developed computational methods which aim at extracting the biological signal of interest from the noisy observations. In this talk, we will focus on probabilistic models designed for this task....

Booking cancellations contribute negatively to the production of accurate forecasts, a critical tool in the hospitality industry. To lessen this influence, hotels implement rigid cancellation policies and overbooking strategies, which in turn can negatively impact revenue and the hotel’s social reputation. To tackle the uncertainty arising from booking cancellations, we combined data from eight hotels’ bookings management systems with data from other sources (weather, holidays,...

We want to guess the position of a target, based only on how weakly or strongly its emitted wireless signal is received at a few places. Assuming usual probabilistic properties of wireless channels, we proceed to frame this problem as a maximum likelihood (ML) estimation problem. The ML estimate, however, is deceptively hard to compute, for it lies in the bowels of a difficult nonconvex...

Genetic testing has long been a major promise to foster personalized healthcare and personalized wellness. The major barrier to its success and massification is the gap between getting the genetic data and providing actionable advice, alongside the high cost of the tests themselves as well as data modelling challenges. Recent trends in genetic testing are showing strong signals that the adoption curve is about to grow...

Wikipedia is the biggest encyclopedia ever created and the fifth most visited website in the world. Tens of millions of people surf it every day, seeking answers to various questions. Collective user activity on its pages leaves publicly available footprints of human behavior, making Wikipedia an excellent source for analysis of collective behavior. Collective memory is an interesting social phenomenon of human behavior. Studying this...

We will present an ongoing project at Priberam that covers aspects that go from the definition of ontologies, corpus gathering and annotation criteria to the concrete annotation task, and finally the development of an engine that automatically extracts, labels and enables the exploration of textual data obtained from medical records. We will report on the resulting nested named-entity recognizer based on stack-LSTMs that achieved state-of-the-art...

The seminar will be a brief overview of deep learning techniques for semantic segmentation of images. Then, it will show a use case of deep learning usage for semantic segmentation of ischemic stroke lesions using 3D MRI data....

In this talk I will introduce a combination of moment based predictive models with deep reinforcement learning architectures, Recurrent Predictive State Policy (RPSP) networks. Predictive state serves as an equivalent representation of a belief state. Therefore, the policy component of the RPSP-network can be purely reactive, simplifying training while still allowing optimal behaviour. We show the efficacy of RPSP-networks under partial observability on a set...

Since the first perceptron was implemented in an IBM 7042 to the success of Deep Blue against a human opponent and, more recently, IBM Watson's victory on Jeopardy!, IBM has been intimately involved with the advancement of cognitive systems. With IBM Research exploring artificial intelligence and machine learning for decades the results are not limited to crucial theoretical advancements but exist as ready-to-use solutions that...