Due to the availability of various open source Machine Learning (ML) tools and libraries, developers nowadays can easily implement their purposes by just invoking machine learning APIs without knowing the details of the algorithm. However, the owners of ML tools and libraries usually pay more attention to the correctness and functionality of their algorithm, while spending much less effort on maintaining their code and keeping their code at a high quality level. Considering the popularity of machine learning in today's world, low quality ML tools and libraries can have a huge impact on the software products that use ML algorithms. So in this paper, we conduct an empirical study on real machine learning bugs to examine their patterns and how they evolve over time. We collect three popular machine learning projects on Github, and manually analyzed 329 closed bugs from the perspectives of their bug category, fix pattern, fix scale, fix duration, and type of software maintenance. The results show that (1) there are seven categories of bugs in machine learning programs; (2) twelve different fix patterns are commonly used to fix the bugs; (3) 63.83% of the patches belong to micro-scale-fix and small-scale-fix, and 68.39% of the bugs are fixed within one month; (4) 47.77% of the bug fixes belong to corrective activity from the view of software maintenance.

With the widespread growth of mass content, information-centric networks (ICN) have become one of the research hotspots of future network architecture. One of the important features of ICN is ubiquitous in-network caching. In recent years, the explosive growth of mobile devices has brought content dynamics, which poses a new challenge to the original ICN caching mechanism. This paper focuses on the WiFi mobile scenario of ICN. We design a new path-associated active caching scheme to shorten the time delay of users obtaining content to enhance the user experience. In this article, based on the WiFi scenario, we first propose a solution for neighbor access point selection from a theoretical perspective, considering the caching cost and transition probability. The cache content is then forwarded based on the selected neighbor set. For cached content, we propose content freshness according to mobile characteristics and consider content popularity at the same time. For cache nodes, we focus on the size of the remaining cache space and the number of hops from the cache to the user. We have implemented this strategy based on the value of caching on the forwarding path. The simulation results show that our caching strategy has a significant improvement in performance compared with active caching and other caching strategies.