Reading your mind, from your EEG!

New research from Mindsee partners [Aalto University, the Helsinki Institute for Information Technology (HIIT) and the Centre of Excellence in Computational Inference (COIN)] reports that for the first time, information retrieval is possible with the help of EEG interpreted with machine learning. Or in other words, Mindsee reading minds!

http://cs.aalto.fi/en/current/news/2016-12-08-002/

 

New Frontiers in Neuroscience article from MINDSEE partner TU Berlin

We're happy to share news of a new Frontiers in Neuroscience article from Mindsee project partner Technical University of Berlin (link below opens in new window):

http://journal.frontiersin.org/article/10.3389/fnins.2016.00530/full

Prof Benjamin Blankertz, PI on MINDSEE at TUB, said "We are pleased to announce a
new review paper about non-medical applications of BCI technology. Section 6 of the paper discusses MindSee HCI applications."

The Rise of Perceptual Technologies

In his two Intel blog articles (links provided below) the author, Bob Duffy, (2015, 2016) focuses on the shift from implicit to explicit computing that is happening in both the general computing area and in technology companies. This brief blog post emphasizes the growing interest for implicit systems, referred to here as “perceptual technologies”. Duffy points out that for decades, technology has worked on developing algorithms and tools to track precise input and information while humans were the ones doing the thinking and the perceiving. The shift is now towards creating machines capable of precise measurement but also to able to determine the relative state of a situation and use this information to trigger an action.

The posts then delve into the possibility of improving machine learning by allowing the computer to experience how we see the world. When interacting humans implicitly evaluate others’ behaviour to be able to understand their emotions and intentions.
For instance, we may observe the listener’s posture or eye gaze to infer how interested they are in the conversation, but how could this be implemented in machine learning?
Duffy provides a simple example to explain this concept such as the idea of having an interactive screen (fish tank) with a series of targets (fish) that will react to specific people’s actions. If the person is getting closer to the screen the fish will come closer to the surface but, if the individual gets too close the fish will retreat, additionally, if people pass too quickly in front of the screen the fish won’t appear. The implementation of these types of applications enable the machine learning process and will aid the computer in understanding how we see the world ultimately making the experience more intuitive and more human. 


Duffy also pictures a future in which these implicit technologies will be embedded in many parts of our life. We all saw glimpses of this in movies such as Minority Report (2002) in which the main character enters a room and is targeted by personal advertising; and this notion is already not too far removed from our present daily reality.  
We can also see it in our environment, it won’t be long until our devices (e.g., car) will be able to sense and identify both our presence and intent and then respond with a congruent action (e.g., parking). Technology is, literally, opening doors and differently from the past developers already have the tools to build these type of systems and the technologies at the base of this area of computing are the internet of things, perceptual technologies and cloud computing.

                                                                  Intel® RealSense™ Technology

                                                                  Intel® RealSense™ Technology

Since 2013 Mindsee has been working towards the goal of developing a system able to understand the implicit state of the user engaged in a scientific literature task; and Bob Duffy’s post show how relevant perceptual technologies have become and how important applications like Mindsee will be in the close future

Source: 

  • Duffy, B (2015) - https://blogs.intel.com/evangelists/2015/07/22/developers-need-think-relatively-intel-realsense
  • Duffy, B (2016) - https://blogs.intel.com/evangelists/2016/04/13/the-shift-from-explicit-to-implicit-computing/ 

 

A Visual History of Human Knowledge

In this TED talk, infographics expert Manuel Lima explores the history of information visualisation. Since the beginning of times humans had the urge to rank, create an order and visualise information. Originally people used the metaphor of a tree to map family relationships, animal species or the realms of human knowledge. However knowledge cannot always be represented in such way: it is too complex, intricate and interdependent. 

The metaphor that better represents knowledge is not a tree but a matrix, a network or a web. Hierarchy can show you the internal structure of an organisation, while a network is able to show the relationships and the connections between its members. Network visualisation is becoming a syntax of a new language and the same matrix model, such as the radial converge, can be use to represent the most disparate things: from DNA to Facebook networks.

Matrices and network seems to be able to embody the notion of decentralisation, non linearity, interconnectedness, interdependence and multiplicity. Manuel Lima believes that this way of thinking will help us solving a number o problems: from decoding the brain to understanding the universe.

Source https://www.youtube.com/watch?v=BQZKs75RMqM&feature=em-uploademail

 

MindSee partners at the 37th Annual International Conference of the IEEE Engineering in Medicine and Biology Society

The 37th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBS) took place on August 25-29, 2015 in Milan, Italy. The conference covers many topics, from Biomedical Signal Processing to Neural and Rehabilitation Engineering, Biomechanics, from Robotics to Biosensors and Bio-Micro/Nano Technologies.  The conference remarks the central role of BME in the improvement and innovation of health care (with a direct impact on the quality of life) but also focuses on how to reach and maintain a “wellness” through proper and advanced technologies, devices and protocols.

Three papers were accepted and presented at the conference, resulting from work conducted by MindSee partners at Technische Universität Berlin.

Papers:

The full program and papers are available here http://emb.citengine.com/event/embc-2015/details.

38th International ACM SIGIR Conference

MindSee related work was presented this August (9th - 13th) at the 38th International ACM SIGIR Conference in Santiago, Chile. SIGIR 2015 is a perfect theatre for Mindsee as it is a major international forum for the presentation of new research results and for the demonstration of new systems and techniques in Information Retrieval.

Mindsee partners at Aalto University presented their work at the main conference on Tuesday the 11th (SciNet: Interactive intent modeling for information discovery) as well as the at the conference's workshop (NeuroIR), focusing on Neuro-Physiological Methods in Information Retrieval Research. A paper (Predicting relevance of text from neuro-physiology ) as well as a demo session were delivered.

6th MindSee Consortium meeting & Exploitation/Innovation session

The next consortium meeting will see all partners from the MindSee project gathering in the town of Helsinki for a 2 day meeting on June 15th and 16th, 2015. Pia Erkinheimo from Digile,  will join the Exploitation and Innovation session that the consortium has organised for day 2, as an external expert and moderator to help the MindSee project with appropriate planning.

See you all in Helsinki!

Read More