Introduction
Technology communities vary depending on their location and subject of interest. Still, attending a meeting where some specific topics related to technology are discussed can be an interesting experience. Meetup groups are an excellent way to find an event that can enhance knowledge and allow meeting new people who share the same interests. For this paper, a meetup dedicated to object detection with deep convolutional networks in images organized by the group named Machine Learning and Artificial Intelligence was attended.
Technology community meeting
The meeting took place at the Innovation Center at Bayview Yards in Ottawa. The agenda of the meeting included arrival and networking with some snacks available, the talk of the speaker that lasted for around forty minutes, and post discussion and talk networking. The group that convened the meeting gathers regularly for monthly meetups. This group observes the technologies related to robotics issues.
It is also noted that “machine learning seeks to develop methods for computers to improve their performance at certain tasks based on observed data” (Ghahramani, 2015, 452). Artificial intelligence and machine learning are widely used for developing drones, diverse autonomous vehicles, and computer bots. They gather at the same place every time, which makes it easy to find and join them. The setting for the meeting is ordinary and reminds most of the lecture settings.
For the announced meeting, a professor at the School of Electrical Engineering and Computer Science was invited. He was the main speaker, and the participants were able to ask him questions after his speech. The purpose of this meeting was to discuss applications and concepts of object detection by using deep convolutional networks. It is a general opinion that this technology has revolutionized the study in computer vision, which resulted in significant improvements in performance.
Some researchers state that the availability of data of large scale for training and convolutional neural networks caused this rapid change (He, Zhang, Ren, & Sun, 2015). It is also emphasized that such approaches as “Region Proposal Network (RPN) takes an image (of any size) as input and outputs a set of rectangular object proposals, each with an objectness score” (Ren, He, Girshick, & Sun, 2017, p. 1138).
The speech of the professor was focused on the problems of image classification and basic concepts of convolutional networks. The speaker described the major strategies applied to build deep architectures that can decline the level of object location used in images. The problems of pedestrian detection by autonomous vehicles and computer monitoring of purchasers in supermarkets were presented. This discussion was interesting and gave a more clear understanding of recent studies dedicated to object detection. For example, in Chapter 10, Zeiler and Fergus (2014) note that convolutional networks demonstrate good performance at such tasks as face detection and digital classification of hand-writing.
After the speech, everyone was able to take part in the discussion and ask questions. The audience amounted to about two hundred people, and the discussion was rather long and interesting. The attenders mainly consisted of white men, but there were quite many women as well. The age of the participants was from twenty to thirty approximately. It was possible to talk to other individuals during the discussion and share one’s standpoint.
Conclusion
This type of gathering was a rather interesting activity to participate in. It was possible to get a deeper insight into the problems and challenges of artificial intelligence and the technology of object detection by machines. The obtained information enhanced the knowledge received during the classes.
References
Ghahramani, Z. (2015). Probabilistic machine learning and artificial intelligence. Nature, 521(7553), 452.
He, K., Zhang, X., Ren, S., & Sun, J. (2015). Spatial pyramid pooling in deep convolutional networks for visual recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence, 37(9), 1904-1916.
Ren, S., He, K., Girshick, R., & Sun, J. (2017). Faster R-CNN: Towards real-time object detection with region proposal networks. IEEE Transactions on Pattern Analysis and Machine Intelligence, 39(6), 1137-1149.
Zeiler, M. D., & Fergus, R. (2014). Visualizing and understanding convolutional networks. In D. Fleet, T. Pajdla, B. Schiele, & T. Tuytelaars (Eds.), European Conference on Computer Vision (pp. 818-833). Toronto, Canada: Springer.