The program in pdf. Poster-presenters: please email a single slide pdf for poster spotlight to Alessadro Ialongo, firstname.lastname@example.org We need a summary slide, useful for a 1 minute spotlight, not a text-heavy description.
The morning run ernsi tradition continues: meet at main gate/porters lodge at 7:00
The twenty-seventh ERNSI Workshop on System Identification is an invitation-only event, to be held September 23 – September 26, 2018 at Pembroke College, Cambridge, U.K. The workshop format will be as usual, with arrival on Sunday afternoon/evening, the program starts on Monday morning and ends after lunch on Wednesday.
Natural Robotics Lab, Department of Automatic Control and Systems Engineering, University of Sheffield. Website: http://www.shef.ac.uk/acse/staff/roderich-gross
In this talk, we study the behavior of autonomous agents, with particular emphasis on systems comprising of numerous embodied agents (e.g., swarms of robots). First, we consider the problem of designing behavioral rules for robots of extreme simplicity. We show among others how “computation-free” robots, with only 1 bit or trit of sensory information, can accomplish tasks such as multi-agent rendezvous, cooperative object manipulation and collective choice. Second, we consider the problem of inferring the behavioral rules or morphology of robots. We use Turing Learning - a generalization of Generative Adversarial Networks. In Turing Learning, the discriminators are allowed to “interrogate”, similar to their human counterparts in the Turing test. We present two case studies where this active learning approach helps improve model accuracy, and discuss applications to robotics and beyond.
Signal Processing and Communications Group, Department of Engineering, University of Cambridge. Website: http://www2.eng.cam.ac.uk/~ik355
One of the main obstacles in the development of effective algorithms for inference and learning from discrete time series date, is the difficulty encountered in the identification of useful temporal structure in the data. We will discuss a class of novel methodological tools for effective Bayesian inference and model selection for general discrete time series, which offer promising results on both small and big data. Our starting point is the development of a rich class of Bayesian hierarchical models for variable-memory Markov chains. The particular prior structure we adopt makes it possible to design effective, linear-time algorithms that can compute most of the important features of the resulting posterior and predictive distributions without resorting to MCMC. We have applied the resulting tools to numerous application-specific tasks, including on-line prediction, segmentation, classification, anomaly detection, entropy estimation, and causality testing, on data sets from different areas of application, including data compression, neuroscience, finance, genetics, and animal communication. Results on both simulated and real data will be presented.
The cost per delegate will be £200 for the event (which includes meals), and £35 (incl. VAT) per night for college accommodation in Pembroke College. The conference fee (£200) will be invoiced from us collectively to each team, whereas the accommodation (£35 per night) should be paid directly to the college at the meeting.
Train: Cambridge is 45 minutes by train from London Kings Cross Station. the EuroStar arrives at St Pancras Station, right across the street from Kings Cross.
Air: Cambridge is 30 minutes by train from London Stansted Airport. Heathrow and Gatwick airports are about 2 hrs away. Other possibilities include London City Airport and Luton Airport.
Talks: All speaking slots (except for invited talks) are 30 minutes long. Speakers are requested to prepare to talk for about 20 minutes.
Posters: Posters should be max A1 and preferably in portrait mode. There will also be a 1 minute spotlight for each poster, we will require a single page pdf for this purpose.
Co-chair: Martin Enqvist