以“人工智能改变世界”为主题的2017中国计算机大会于10月26日至28日在福州举办。中国计算机大会由中国计算机学会创建于2003年，已成功举办十三届，成为国内规格最高、规模最大的计算领域的盛会。“人工智能”近几年热度高涨，无论是学术界还是产业界、投资界都高度聚焦，公众对人工智能的期待也在不断提升。但是支撑人工智能发展的还是算法，因此如何让科学、技术与应用相结合，是中国计算机学会一直关注并致力推进的。
本次大会安排了30场论坛，涉及人工智能、类脑计算、量子计算、区块链、智能存储、深度学习、智慧城市、健康医疗大数据等前沿技术话题。300余名国内外计算机领域知名专家、企业家到会演讲，参会者达6000余人。
浙大求是高等研究院系统神经与认知研究所应邀参加了会议并做了大会报告。（内容如下）
Developing a network theory of brain function via a columnar connectome.
To understand the principles underlying brain function, we must develop a systematic understanding of how information is routed from one area to another. What are the mathematics of 2D-2D transforms? What are the possible routes, the hubs, the basic topologies. How many processing stages are necessary and sufficient between sensory input stage and final output stage? How does brain function change when information is channeled through a different path or when normal paths are not available such as in disease? Our goal is to collect extensive brain network datasets and then develop a topological understanding of these networks. In this way, we may arrive at a mathematics of brain function.
The monkey connectome data we are collecting will provide multiple examples of types of inter-areal and intra-areal connectivity patterns. We will use the word 'network' to refer to these patterns. Network parameters to be examined will include: size of cortical area (number of total modules in area), number of module types (e.g. color, orientation), number of modules in network, geometry of the network (e.g. anisotropy, topology), convergence and divergence between areas. 'Small world' and graph theory approaches will be used. Particular attention will be paid to whether there are canonical (from a topological viewpoint) connection patterns across topographic location (e.g. central to peripheral vision), sensory modalities (e.g. vision, touch) and across hierarchical levels (e.g. early visual vs higher order visual areas). One compelling question is whether there are only a few topologies that explain all connection patterns. If so, this approach could make predictions for areas whose functional organizations are not understood.
