ABSTRACT Private set operations (PSO) address key challenges in cross‐organizational data sharing where multiple transport entities securely verify shared cargo manifests or delivery schedules without exposing sensitive information. Private set intersection (PSI) enables secure computation of common elements across private datasets, while its variant, PSI with cardinality (PSI‐CA), reveals only the intersection size. The PSO protocol based on Diffie‐Hellman (DH) key agreement has emerged as a prominent solution for communication‐sensitive applications. However, current research efforts mostly concentrate on two‐party implementations, which face inherent limitations in addressing the communication complexity when scaling to multiple participants. In this work, we firstly present E3PSI‐CA, a novel three‐party PSO protocol. This construction for the semi‐honest model leverages bilinear maps in DH key agreement and Bloom Filters (BF) for compact set representation, achieving remarkable efficiency. By integrating private information retrieval (PIR) techniques, the party in E3PSI‐CAcan efficiently compute the set intersection. To address the inherent accuracy limitations of the BF and extend security to the malicious model, we further propose an enhanced PSI scheme. This variant is built upon a “Ring” structure, within which we employ DH key agreement and OKVS data structure to achieve security against any single malicious party, while the OKVS simultaneously ensures perfect accuracy (zero false positives). We prove the security through the ideal/real simulation paradigm and conduct performance evaluations.

The rapid advancement of the Internet has brought the exploration and management of the IPv6 address space to the forefront of network research. With its 128-bit addressing scheme, IPv6 provides approximately 340 undecillion addresses, effectively mitigating the address exhaustion issue inherent in IPv4. The immense address space of IPv6 renders traditional brute-force scanning, once effective for IPv4, no longer feasible. This challenge calls for the development of efficient and precise IPv6 address generation methods to ensure the stability and security of the Internet. To tackle this challenge, this study introduces an innovative IPv6 address generation method leveraging a fine-tuned Large Language Model (LLM), enhanced by the vLLM inference framework to improve generation efficiency. This approach leverages the strong learning and generalization capabilities of LLMs to effectively handle the complexities of IPv6 address generation. We begin with an initial classification of IPv6 addresses to reduce the model’s learning complexity. Next, we curate a fine-tuning dataset and apply the LoRA technique to fine-tune the Qwen-7B-chat model. The results show that the fine-tuned LLM outperforms existing methods in generating IPv6 addresses, especially in terms of generation speed and accuracy.

Multi-party private set intersection (PSI) enables multiple parties to compute the common items of private sets without disclosing any other information beyond the result; thus, it has gained significant importance in various distributed computation scenarios. As the number of participants increases, the performance of multi-party PSI protocols is significantly influenced, primarily by the number of interaction rounds needed. In this study, we propose two novel multi-party PSI protocols: the first one is 1MPSI, and the second is 2MPSI. The 1MPSI appears as a wheel structure where the parties need only one round of interactive communication online. 1MPSI is based on the Ring version of Oblivious Linear-function Evaluation (OLE). Benefiting from the wheel structure, 1MPSI supports parallel computation and achieves competitive efficiency between the leader and the other participants (OLE receivers) after input-independent precomputation. The 2MPSI adopts a dual-core star structure and introduces Oblivious Key-Value Store (OKVS), which results in better performance when handling larger set sizes and more participants. Our protocols are designed with simplicity and ease of implementation in mind. Experimental evaluations demonstrate the superiority of our protocols over current open-source multi-party PSI protocols as the set size increases from 212to 220when involving 10 and 16 parties. In a test with 16 parties each inputting 220elements, 2MPSI achieves a runtime of only 65 seconds.

Aiming at the problem that conventional intrusion detection system is difficult to defend against Crossfire attacks, a method of detecting and defending Crossfire attacks is designed after analyzing the characteristics of Crossfire attacks deeply. This method utilizes Software Defined Network (SDN). Firstly, vulnerable network bottleneck nodes and links are screened out. On this basis, virtual topology is introduced to prevent Crossfire attacks. Finally, a slow start detection defense strategy is adopted to prevent Crossfire attacks. Experimental results show that the proposed method can effectively detect and mitigate such attacks in the SDN environment, and the normal forwarding of legitimate traffic in the physical topology is basically not affected in this process.

Due to the lack of face-to-face interactions, online learners frequently experience social isolation that negatively impacts students’ well-being and learning experiences. Many text-based AI agents have been equipped with different social characteristics and functionalities to support people who are socially isolated. However, the design of agent’s functionalities, social characteristics, and ethical challenges in promoting social connectedness among online learners are underexplored. Taking a co-design approach, we included 23 online learners enrolled in an online for-degree graduate program as active participants in two virtual co-design workshop studies. Through four different co-design activities, we identified online learners’ preferences for AI agent’s functionalities and social characteristics in promoting their social connectedness as well as potential ethical concerns. Based on our findings, we establish the role of AI agent as a facilitator to continuously scaffold online learners’ social connection process. We further discuss the unique ethical challenges regarding agent-mediated social interaction in online learning.

Now various privacy-preserving techniques have been combined with machine learning to ensure training data security. However, during the training, users were unable to select data and labels adaptively in the server, which made the trained models challenging in meeting user needs. Repeated model training not only wastes server resources, but also reduces query efficiency. In this paper, we combine two privacy-preserving technologies, searchable encryption and homomorphic encryption, and propose a highly flexible machine training framework. Homomorphic encryption technology is used to encrypt data, and symmetric searchable encryption technology is used to generate blind indexes and trapdoors. The server can perform ciphertext data search through blind query trapdoors and indexes, and can use the searched homomorphic ciphertext for model training. This framework significantly improves the flexibility of training and the fit of the model, and also ensures the security of dynamic data management. The server locally uses the header of the query trap and the storage address of the ciphertext model to build an automatically updatable ciphertext model table. On the basis of ensuring the effectiveness of the model, it greatly improves the efficiency of obtaining the model, effectively prevents redundant training when different users send the same request, and reasonably allocates resources in the server.

Our online interactions are constantly mediated through Artificial Intelligence (AI), especially our social interactions. AI-mediated social interaction is the AI-facilitated process of building and maintaining social connections between individuals through information inferred from people's online posts. With its impending application across a number of contexts, the challenges and opportunities of AI-mediated social interaction remain underexplored. This paper seeks to understand the design space of AI-mediated social interaction in the context of online learning, where students frequently face social isolation. We deployed an AI agent named SAMI in three class discussion forums to help online learners build social connections. Using SAMI as a probe, we conducted semi-structured interviews with 26 students to understand their difficulties in remote social interactions and their experiences with SAMI. Through the lenses of social translucence and social-technical gap, we illustrate online learners' difficulties in remote social interactions and how SAMI resolved some of the difficulties. We also identify potential ethical and social challenges of SAMI such as user agency and privacy. Based on our findings, we outline the design space of AI-mediated social interaction. We discuss the design tension between AI performance and ethical design and pinpoint two design opportunities for AI-mediated social interaction in designing towards human-AI collaborative social matching and artificial serendipity.

As online education proliferates, one concern that has been raised is that it may fail to capture desirable emergent phe-nomena from on-campus programs. Student community is one example of such a phenomenon: on-campus student communities thrive based on synchronous collocation. An online program might be designed to capture all deliberate constructs in an on-campus program, but there may be beneficial side effects of synchronous collocation that are not apparent. In this work, we examine the issue of social isolation in an online graduate program. By happenstance, three studies were conducted in relative isolation looking at social isolation from different angles. The first study exam-ined trajectories in social presence as a semester proceeded. The second study developed an understanding of students' needs with regard to community in an online program. The third study tested out an immersive virtual environment to try to improve students' sense of connectedness. Combin-ing their findings, we find compelling evidence of the exist-ence of a Synchronicity Paradox in online education: stu-dents desire synchronicity to form strong social communi-ties, and yet part of the chief appeal of these online pro-grams is their asynchronicity. In light of this finding, we provide design guidelines for how synchronicity may be reintroduced into asynchronous programs without sacrific-ing the benefits of asynchronicity. More specifically, we propose that scale itself may be the key to building emer-gent synchronicity.

Large-scale educational settings have been common domains for affect detection and recognition research. Most research emphasizes improvements in the accuracy of affect measurement to enhance instructors' efficiency in managing large numbers of students. However, these technologies are not designed from students' perspectives, nor designed for students' own usage. To identify the unique design considerations for affect sensors that consider student capacities and challenges, and explore the potential of affect sensors to support students' self-learning, we conducted semi-structured interviews and surveys with both online students and on-campus students enrolled in large in-person classes. Drawing on these studies we: (a) propose using affect data to support students' self-regulated learning behaviors through a "scaling for empowerment'' design perspective, (b) identify design guidelines to mitigate students' concerns regarding the use of affect data at scale, (c) provide design recommendations for the physical design of affect sensors for large educational settings.

Large-scale data training is vital to the generalization performance of deep learning (DL) models. However, collecting data directly is associated with increased risk of privacy disclosure, particularly in special fields such as healthcare, finance, and genomics. To protect training data privacy, collaborative deep learning (CDL) has been proposed to enable joint training from multiple data owners while providing reliable privacy guarantee. However, recent studies have shown that CDL is vulnerable to several attacks that could reveal sensitive information about the original training data. One of the most powerful attacks benefits from the leakage from gradient sharing during collaborative training process. In this study, we present a new CDL framework, PrivateDL, to effectively protect private training data against leakage from gradient sharing. Unlike conventional training process that trains on private data directly, PrivateDL allows effective transfer of relational knowledge from sensitive data to public data in a privacy-preserving way, and enables participants to jointly learn local models based on the public data with noise-preserving labels. This way, PrivateDL establishes a privacy gap between the local models and the private datasets, thereby ensuring privacy against the attacks launched to the local models through gradient sharing. Moreover, we propose a new algorithm called Distributed Aggregation Stochastic Gradient Descent, which is designed to improve the efficiency and accuracy of CDL, especially in the asynchronous training mode. Experimental results demonstrate that PrivateDL preserves data privacy with reasonable performance overhead.