Search

With the development of deep learning, natural language processing technology has effectively improved the efficiency of various aspects of the traditional judicial industry. However, most current efforts focus solely on individual judicial stage, overlooking cross-stage collaboration. As the autonomous agents powered by large language models are becoming increasingly smart and able to make complex decisions in real-world settings, offering new insights for judicial intelligence. In this paper, (1) we introduce SimuCourt, a judicial benchmark that encompasses 420 judgment documents from real-world, spanning the three most common types of judicial cases, and a novel task Judicial Decision-Making to evaluate the judicial analysis and decision-making power of agents. To support this task, we construct a large-scale judicial knowledge base, JudicialKB, with multiple legal knowledge. (2) we propose a novel multi-agent framework, AgentsCourt. Our framework follows the real-world classic court trial process, consisting of court debate simulation, legal information retrieval and judgement refinement to simulate the decision-making of judge. (3) we perform extensive experiments, the results demonstrate that, our framework outperforms the existing advanced methods in various aspects, especially in generating legal grounds, where our model achieves significant improvements of 8.6% and 9.1% F1 score in the first and second instance settings, respectively.

The popular notion of Internet of Things (IoT) implies two salient features: 1. a diversity of small things, i.e., constrained devices; 2. their seamless integration with the Internet. Pioneering work in Wireless Sensor Networks (WSNs) have laid a solid technological foundation for autonomous, low power wireless communication among battery-powered, microcontroller-based devices. On the other hand, as devices are being connected to the Internet in large numbers, industry experts and regulators have associated IoT with enormous security risk. Sensitive personal information, highly complex business workflows, and critical infrastructure for public safety are at stake. In this dissertation, we first explore the scalability of IoT. Approaching from the particular angle of radio interference, we study unstable and faulty network behavior when links between low power radios are disrupted. Our low cost and practical interference generation tools fill a gap between protocol design and test. We then underline the threat of novel attacks at the physical layer, which lead to denial of service and battery draining of low power radios. Launched from low cost hardware, the attacks we devise are power-efficient and hard to detect; and they reach longer ranges than jamming. Finally, we take a step closer to realization of secure and large-scale IoT deployment by enabling certificate enrollment, a key component in a public key infrastructure, for small devices. We show that automated enrollment of device certificates becomes feasible when a memory and power efficient IoT protocol stack is leveraged. Spanning between the physical layer and the application layer, our work has enriched the knowledge domain of IoT and advanced the technological frontier of scalable and secure IoT deployment.

Low-power wireless networks transmit at low output power and are hence susceptible to cross-technology interference. The latter may cause packet loss which may waste scarce energy resources by requiring the retransmission of packets. Jamming attacks are even more harmful than cross-technology interference in that they may totally prevent packet reception and hence disturb or even disrupt applications. Therefore, it is important to recognize such jamming attacks. In this paper, we present JamSense. JamSense extends SpeckSense, a system that is able to detect multiple sources of interference, with the ability to classify jamming attacks. As SpeckSense, JamSense runs on resource-constrained nodes. Our experimental evaluation on real hardware shows that JamSense is able to identify jamming attacks with high accuracy while not classifying Bluetooth or WiFi interference as jamming attacks.

Glossy networks make use of concurrent transmissions to achieve rapid network flooding in wireless networks with high reliability. They are robust against jamming and header injection attacks. We find that Glossy floods can be hijacked by a packet injection attacker to penetrate into the network and cause severe loss. We demonstrate the design of such an attacker by evaluating its effectiveness in a 30-node testbed., Promos

The past couple of years have seen a heightened interest in the Internet of Things (IoT), transcending industry, academia and government. As with new ideas that hold immense potential, the optimism of IoT has also exaggerated the underlying technologies well before they can mature into a sustainable ecosystem. While 6LoWPAN has emerged as a disruptive technology that brings IP capability to networks of resource constrained devices, a suitable radio technology for this device class is still debatable. In the recent past, Bluetooth Low Energy (LE) - a subset of the Bluetooth v4.0 stack -has surfaced as an appealing alternative that provides a low-power and loosely coupled mechanism for sensor data collection with ubiquitous units (e.g., smartphones and tablets). When Bluetooth 4.0 was first released, it was not targeted for IP-connected devices but for communication between two neighboring peers. However, the latest release of Bluetooth 4.2 offers features that makes Bluetooth LE a competitive candidate among the available low-power communication technologies in the IoT space. In this paper, we discuss the novel features of Bluetooth LE and its applicability in 6LoWPAN networks. We also highlight important research questions and pointers for potential improvement for its greater impact.