In recent years, synthesizing drugs powered by artificial intelligence has brought great convenience to society. Since retrosynthetic analysis occupies an essential position in synthetic chemistry, it has received broad attention from researchers. In this review, we comprehensively summarize the development process of retrosynthesis in the context of deep learning. This review covers all aspects of retrosynthesis, including datasets, models and tools. Specifically, we report representative models from academia, in addition to a detailed description of the available and stable platforms in the industry. We also discuss the disadvantages of the existing models and provide potential future trends, so that more abecedarians will quickly understand and participate in the family of retrosynthesis planning.

In recent years, many organizations have established bounty programs that attract white hat hackers who contribute vulnerability reports of web systems. In this paper, we collect publicly available data of two representative web vulnerability discovery ecosystems (Wooyun and HackerOne) and study their characteristics, trajectory, and impact. We find that both ecosystems include large and continuously growing white hat communities which have provided significant contributions to organizations from a wide range of business sectors. We also analyze vulnerability trends, response and resolve behaviors, and reward structures of participating organizations. Our analysis based on the HackerOne dataset reveals that a considerable number of organizations exhibit decreasing trends for reported web vulnerabilities. We further conduct a regression study which shows that monetary incentives have a significantly positive correlation with the number of vulnerabilities reported. Finally, we make recommendations aimed at increasing participation by white hats and organizations in such ecosystems.

Buffer over-read vulnerabilities (e.g., Heartbleed) can lead to serious information leakage and monetary lost. Most of previous approaches focus on buffer overflow (i.e., over-write), which are either infeasible (e.g., canary) or impractical (e.g., bounds checking) in dealing with over-read vulnerabilities. As an emerging type of vulnerability, people need in-depth understanding of buffer over-read: the vulnerability, the security risk and the defense methods. This paper presents a systematic methodology to evaluate the potential risks of unknown buffer over-read vulnerabilities. Specifically, we model the buffer over-read vulnerabilities and focus on the quantification of how much information can be potentially leaked. We perform risk assessment using the RUBiS benchmark which is an auction site prototype modeled after eBay.com. We evaluate the effectiveness and performance of a few mitigation techniques and conduct a quantitative risk measurement study. We find that even simple techniques can achieve significant reduction on information leakage against over-read with reasonable performance penalty. We summarize our experience learned from the study, hoping to facilitate further studies on the over-read vulnerability.

For decades buffer overflows have been one of the most prevalent and dangerous software vulnerabilities. Although many techniques have been proposed to address the problem, they mostly introduce a very high overhead while others assume the availability of a separate system to pinpoint attacks or provide detailed traces for defense generation, which is very slow in itself and requires considerable extra resources. We propose an efficient solution against heap buffer overflows that integrates exploit detection, defense generation, and overflow prevention in a single system, named Heap Therapy. During program execution it conducts on-the-fly lightweight trace collection and exploit detection, and initiates automated diagnosis upon detection to generate defenses in real-time. It can handle both over-write and over-read attacks, such as the recent Heartbleed attack. The system has no false positives, and keeps effective under polymorphic exploits.%as the generated defense captures semantic characteristics of exploits. It is compliant with mainstream hardware and operating systems, and does not rely on specific allocation algorithms. We evaluated Heap Therapy on a variety of services (database, web, and ftp) and benchmarks (SPEC CPU2006), it incurs a very low average overhead in terms of both speed (6.2%) and memory (7.7%).

In a federated database system, each independent party exports some of its data for information sharing. The information sharing in such a system is very inflexible, as all peer parties access the same set of data exported by a party, while the party may want to authorize different peer parties to access different portions of its information. We propose a novel query evaluation scheme that supports differentiated access control with decentralized query processing. Anew efficient join method, named split-join, along with other safe join methods is adopted in the query planning algorithm. The generated query execution reduces the communication cost by pushing partial query computation to data sources in a safe way. The proofs of the correctness and safety of the algorithm are presented. The evaluation demonstrates that the scheme significantly saves the communication cost in a variety of circumstances and settings while enforcing autonomous and differentiated information sharing effectively.