The Machine Learning Summer School (#MLSS2018), held at the Autonomous University of Madrid, has finally come to an end. It was an awesome event and very laudably met the expectations I outlined in my anticipation post. I will be doing a three-part series on the summer school experience beginning with this post. The first poster session was held on the evening of Thursday, 30th of August 2018, after the day’s lectures. It was a relaxed event with over 50 presenters interacting with peers in small groups or one-on-one about their poster. Towards the end, I got the chance to visit other posters I found interesting or related to my research.
My poster abstract as submitted to MLSS
This research proceeds at the edge where the focus is to process video streams so as to detect and isolate special events. This allows the cloud to receive just as much data as required for predictive analytics, pattern recognition and data mining. It also reduces the amount of data stored in the cloud by ensuring that only special events are filtered starting from the edge devices to the cloud. Using an online deep learning algorithm, the anomalous events are captured in real-time and transmitted to the next tier of the network. While there has been a lot of research in individual fields of edge learning, video analytics, cloud data fusion and anomaly detection, research is still lacking in the aggregation of these technologies, where anomalous activities in visual networks can be detected through a hybrid learning method between the edge and the cloud. The hybrid learning real-time analytics could be valuable in various applications like surveillance systems and environmental monitoring.
MLSS website (first session, no. 36)
Interactions and feedback
I don’t remember exactly how many people I interacted with about my poster during the session, but I do remember some conversations that stood out especially because the chats went beyond that evening. Lukas Schott was the guy who asked about what happens in the latent space that would later lead us to meet over lunch to talk more in-depth about ConvNet model for my problem domain. Hady Elsahar was interested in the anomaly detection bit. He made funny remarks but which I considered a possibly useful angle to keep in mind. He said things like: “Hope your anomaly detector wouldn’t be discriminatory”. Also, he asked: “What if abnormal behaviour becomes normal for instance in a crime-prone environment?”. The rest of the summer school, he would introduce me to peers as the anomaly detector.
Hope you found the abstract and poster explicit enough? You are welcome to ask questions or give your feedback in the comment section below. In part 2 and 3 of this summer school series, I will be talking about the social side of the experience and the insights I gathered for my research. So don’t forget to check back soon.
Great experience you got there, keep shinning!