Extracting entities and relations from unstructured text has attracted increasing attention in recent years but remains challenging, due to the intrinsic difficulty identifying overlapping with shared entities. Prior works show that joint learning can result a noticeable performance gain. However, they usually involve sequential interrelated steps suffer problem of exposure bias. At training time, predict ground truth conditions while at inference it make extraction scratch. This discrepancy...

Abstract The rapid development of large language models (LLMs) has opened new avenues across various fields, including cybersecurity, which faces an evolving threat landscape and demand for innovative technologies. Despite initial explorations into the application LLMs in there is a lack comprehensive overview this research area. This paper addresses gap by providing systematic literature review, covering analysis over 300 works, encompassing 25 more than 10 downstream scenarios. Our three...

Social engineering has posed a serious security threat to infrastructure, user, data and operations of cyberspace. Nevertheless, there are many conceptual deficiencies (such as inconsistent intensions, vague boundary, confusing instances, overgeneralization abuse) the term making negative impacts on understanding, analysis defense social attacks. In this paper, an in-depth literature survey is conducted, original meaning in cybersecurity traced, evolution technical development analysed...

Social engineering attacks have posed a serious security threat to cyberspace. However, there is much we yet know regarding what and how lead the success of social attacks. This paper proposes conceptual model which provides an integrative structural perspective describe work. Three core entities (effect mechanism, human vulnerability attack method) are identified help understanding take effect. Then, beyond familiar scope, analyze discuss effect mechanisms involving 6 aspects <italic...

Underwater sensor networks are envisioned to enable a wide range of underwater applications such as pollution monitoring, offshore exploration, and oil/gas spill monitoring. Such require precise location information otherwise the sensed data might be meaningless. On other hand, security privacy critical issues typically deployed in harsh environments. Nevertheless, most localization schemes vulnerable many attacks suffer from potential violations they designed for benign However, scheme that...

Binary code similarity detection is a fundamental technique for many security applications such as vulnerability search, patch analysis, and malware detection. There an increasing need to detect similar search across architectures with the increase of critical vulnerabilities in IoT devices. The variety hardware software platforms requires capture semantic equivalence fragments However, existing approaches are insufficient capturing similarity. We notice that abstract syntax tree (AST)...

Yucheng Wang, Bowen Yu, Hongsong Zhu, Tingwen Liu, Nan Limin Sun. Proceedings of the 59th Annual Meeting Association for Computational Linguistics and 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers). 2021.

Abstract Social engineering has posed a serious threat to cyberspace security. To protect against social attacks, fundamental work is know what constitutes engineering. This paper first develops domain ontology of in cybersecurity and conducts evaluation by its knowledge graph application. The defines 11 concepts core entities that significantly constitute or affect domain, together with 22 kinds relations describing how these related each other. It provides formal explicit schema...

The rapid advancements in large language models (LLMs) have opened new avenues across various fields, including cybersecurity, which faces an ever-evolving threat landscape and need for innovative technologies. Despite initial explorations into the application of LLMs there is a lack comprehensive overview this research area. This paper bridge gap by providing systematic literature review, encompassing analysis over 180 works, spanning 25 more than 10 downstream scenarios. Our addresses...

Concolic execution follows the paths of concrete inputs, capable generating new inputs for unexplored code by solving negated path constraints. However, implicit flows can hinder concolic execution, reducing coverage. Implicit occur when influence control flow, and flow variation affects values some variables. During preceding selections limit potential these This limitation may result in unsolvable constraints, subsequently restricting generation paths. Our insight is that following same...

Abstract A reasonable construction of hollow structures to obtain high‐performance absorbers is widely studied, but it still a challenge select suitable materials improve the low‐frequency attenuation performance. Here, Fe 3 O 4 @C@NiO nanoprisms with unique tip shapes, asymmetric multi‐path cavity, and core–shell heteroepitaxy structure are designed synthesized based on anisotropy intrinsic physical characteristics. Impressively, by changing load NiO, composites achieve strong absorption,...

Federated learning is an essential distributed model training technique. However, threats such as gradient inversion attacks and poisoning pose significant risks to the privacy of data correctness. We propose a novel approach called SMTFL achieve secure in federated without relying on trusted participants. To safeguard gradients against attacks, clients are dynamically grouped, allowing one client's be divided obfuscate other within group. This method incorporates checks balances reduce...

The Internet of Things (IoT) has been considered as the third revolution in digital technology after computer and Internet, it brings significant advantages to smart grid. Disaster prediction prevention power-lines is one most difficult problems for electricity transmission companies. Advanced sensing communication technologies IoT can effectively avoid or reduce damage natural disasters power-lines, hence improve reliability stability power transmission. In this paper, a three-layer...

Clock synchronization is critical for Wireless Sensor Networks (WSNs) due to the need of inter-node coordination and collaborative information processing. Although many message passing protocols can achieve satisfactory clock accuracy, they incur prohibitively high overhead when network scales more than tens nodes. An alternative approach take advantage global time reference induced by existing infrastructures including GPS, timekeeping radio stations, or power grid. However, consumption...

Greybox fuzzing has become one of the most effective vulnerability discovery techniques. However, greybox techniques cannot be directly applied to applications in IoT devices. The main reason is that executing these highly relies on specific system environments and hardware. To execute Linux-based devices, existing use full-system emulation for purpose maximizing compatibility. compared with user-mode emulation, suffersfrom great overhead. Therefore, some previous works, such as Firm-AFL,...

Although the importance of using static taint analysis to detect taint-style vulnerabilities in Linux-based embedded firmware is widely recognized, existing approaches are plagued by following major limitations: (a) Existing works cannot properly handle indirect call on path from attacker-controlled sources security-sensitive sinks, resulting lots false negatives. (b) They employ heuristics identify mediate source and it not accurate enough, which leads high positives.

In the field of Cyber Physical Systems and Pervasive Computing, physical resources web can be easily handled seamlessly integrated into our life. However, due to heterogeneity devices tight coupling individual information systems, developers cannot create their specific applications by combining with resources. this paper, we proposed Rest Thing which is a restful service infrastructure based on REST principles in order hide provide seamless way integrate embedded existing applications....

In recent years, the capacity of cellular networks has hardly kept pace with dramatic explosion traffic from emerging mobile applications. To address this issue, a number studies have been proposed to offload using delay-tolerant (DTNs) opportunistically formed by devices. These off-loading schemes rely on assumption small-size content, which can be entirely transmitted over single DTN contact. However, such an assumption, was probably valid in past, no longer hold increasing popularity...