随着科学、技术、商业、国防、电信、医学、金融和其他行业中海量数据的涌现，人类日益期望从大数据中发现新的科学规律和新的商机。在数据挖掘和机器学习等方法变得更加时髦的同时，他们的内在局限性也变得日益明显。为克服这个局限性，研究者已日益关注用复杂性科学的方法分析大数据的可能性。通过金融危机预测，中国市场经济评估（包括近期股市的崩盘），舆情传播及cyber-influence的刻画，生物医学数据分析，和全球政治冲突和国际关系演化的分析，我们展示用复杂性科学分析大数据的的巨大前景。

Broad applications of multiscale chaotic dynamics in engineering

Engineering has been going to ever larger and smaller scales. As a result, complex data have been accumulating rapidly in almost every area of science and technology. To effectively deal with these data, we discuss a scale-dependent Lyapunov exponent (SDLE) based multiscale approach for characterizing chaotic dynamics. SDLE can characterize all known types of signal data, including deterministic chaos, noisy chaos, random fractal signals, and even intermittent chaos.  As example applications, we discuss characterization of  the chaotic dynamics of semiconductor lasers, circuits with Boolean chaos, biomedical signals, and sea clutter radar returns.

Versatile adaptive multiscale analysis of complex time series

Emergent behaviors of complex systems have fascinated mankind for aeons. It is only in recent decades that extensive efforts have been made to quantitatively study them, resulting in important theories and tools such as chaos theory, random fractal theory, and multiscale analyses. This talk aims to convey some of the best practices in this vast field, emphasizing theory meets reality. In particular, a versatile adaptive filter, which can accurately determine trend, reduce noise, perform fractal and multifractal analysis and multiscale decomposition, and process images, will be presented.