Large Language Models (LLMs), constrained by limited context windows, often face significant performance degradation when reasoning over long contexts.To address this, Retrieval-Augmented Generation (RAG) retrieves and reasons over chunks but frequently sacrifices logical coherence due to its reliance on similarity-based rankings.Similarly, divideand-conquer frameworks (DCF) split documents into small chunks for independent reasoning and aggregation.While effective for local reasoning, DCF struggles to capture longrange dependencies and risks inducing conflicts by processing chunks in isolation.To overcome these limitations, we propose ToM, a novel Tree-oriented MapReduce framework for long-context reasoning.ToM leverages the inherent hierarchical structure of long documents (e.g., main headings and subheadings) by constructing a DocTree through hierarchical semantic parsing and performing bottom-up aggregation.Using a Tree MapReduce approach, ToM enables recursive reasoning: in the Map step, rationales are generated at child nodes; in the Reduce step, these rationales are aggregated across sibling nodes to resolve conflicts or reach consensus at parent nodes.Experimental results on 70B+ LLMs show that ToM significantly outperforms existing divide-andconquer frameworks and retrieval-augmented generation methods, achieving better logical coherence and long-context reasoning.Our code is available at https://github.com/gjn12- 31/ToM.

Large language models (LLMs) have achieved remarkable success across various domains, driving significant technological advancements and innovations.Despite the rapid growth in model scale and capability, systematic, data-driven research on how structural configurations affect performance remains scarce.To address this gap, we present a large-scale dataset encompassing diverse open-source LLM structures and their performance across multiple benchmarks.Leveraging this dataset, we conduct a systematic, data mining-driven analysis to validate and quantify the relationship between structural configurations and performance.Our study begins with a review of the historical development of LLMs and an exploration of potential future trends.We then analyze how various structural choices impact performance across benchmarks and further corroborate our findings using mechanistic interpretability techniques.By providing data-driven insights into LLM optimization, our work aims to guide the targeted development and application of future models.We release our dataset at

Abstract Wireless sensor networks (WSNs) can provide the world with a technology for real‐time event monitoring. One of the primary concerns that hinder the successful deployment of WSNs is source‐location privacy (SLP). The privacy of the source location is vital and highly jeopardized by the usage of wireless communications. Although message content privacy can be ensured through message encryption, it is much more difficult to adequately address the SLP. For WSNs, SLP service is further complex by the fact that sensors consist of low‐cost and energy‐efficient radio devices. Therefore, using computationally intensive cryptographic algorithms (such as public‐key cryptosystems) and large‐scale broadcasting‐based protocols are not suitable for WSNs. In this paper, we analyze the quantitatively measure source‐location information leakage in routing‐based SLP protection schemes for WSNs. Through this model, we identify vulnerabilities of some well‐known SLP protection schemes. We also propose a routing technique, called the Sink Toroidal Region (STaR), to provide adequate SLP with low energy consumption. With this routing technique, the source node randomly selects an intermediate node within a designed STaR area located around the sink node. Furthermore, this routing protocol ensures that the intermediate node is neither too close, nor too far from the sink node in relations to the entire network. While ensuring SLP, our simulation results show that the proposed scheme is very efficient and can be used for practical applications. Copyright © 2012 John Wiley & Sons, Ltd.

Abstract Communication anonymity is becoming an increasingly important, or even indispensable, security requirement for many applications. The existing research in anonymous communications can largely be divided into two categories: mix‐based systems and secure multiparty computation‐based systems, originating from mixnet and DC‐net, respectively. However, they either cannot provide provable anonymity or suffer from transmission collision problem. In this paper, we first propose a novel unconditionally secure source anonymous message authentication code that can be applied to any messages without relying on any trusted third parties. While ensuring message sender anonymity, secure source anonymous message authentication code can also provide message content authenticity. We then propose a novel communication protocol that can hide the senders and the recipients from each other and thus can be used for secure file sharing. The security analysis demonstrates that the proposed protocol is secure against various attacks. Our analysis also shows that it is efficient and practical. Copyright © 2012 John Wiley & Sons, Ltd.

Message authentication is one of the most effective ways to thwart unauthorized and corrupted messages from being forwarded in wireless sensor networks (WSNs). For this reason, many message authentication schemes have been developed, based on either symmetric-key cryptosystems or public-key cryptosystems. Most of them, however, have the limitations of high computational and communication overhead in addition to lack of scalability and resilience to node compromise attacks. To address these issues, a polynomial-based scheme was recently introduced. However, this scheme and its extensions all have the weakness of a built-in threshold determined by the degree of the polynomial: when the number of messages transmitted is larger than this threshold, the adversary can fully recover the polynomial. In this paper, we propose a scalable authentication scheme based on elliptic curve cryptography (ECC). While enabling intermediate nodes authentication, our proposed scheme allows any node to transmit an unlimited number of messages without suffering the threshold problem. In addition, our scheme can also provide message source privacy. Both theoretical analysis and simulation results demonstrate that our proposed scheme is more efficient than the polynomial-based approach in terms of computational and communication overhead under comparable security levels while providing message source privacy.

Message authentication is one of the most effective ways to thwart unauthorized and corrupted traffic from being forwarded in wireless sensor networks (WSNs). To provide this service, a polynomial-based scheme was recently introduced. However, this scheme and its extensions all have the weakness of a built-in threshold determined by the degree of the polynomial: when the number of messages transmitted is larger than this threshold, the adversary can fully recover the polynomial. In this paper, we propose a scalable authentication scheme based on elliptic curve cryptography (ECC). While enabling intermediate node authentication, our proposed scheme allows any node to transmit an unlimited number of messages without suffering the threshold problem. In addition, our scheme can also provide message source privacy. Both theoretical analysis and simulation results demonstrate that our proposed scheme is more efficient than the polynomial-based approach in terms of communication and computational overhead under comparable security levels while providing message source privacy.

Wireless sensor networks (WSNs) have been widely used in many areas for critical infrastructure monitoring and information collection. While confidentiality of the message can be ensured through content encryption, it is much more difficult to adequately address source-location privacy (SLP). For WSNs, SLP service is further complicated by the nature that the sensor nodes generally consist of low-cost and low-power radio devices. Computationally intensive cryptographic algorithms (such as public-key cryptosystems), and large scale broadcasting-based protocols may not be suitable. In this paper, we first propose criteria to quantitatively measure source-location information leakage in routing-based SLP protection schemes for WSNs. Through this model, we identify vulnerabilities of some well-known SLP protection schemes. We then propose a scheme to provide SLP through routing to a randomly selected intermediate node (RSIN) and a network mixing ring (NMR). Our security analysis, based on the proposed criteria, shows that the proposed scheme can provide excellent SLP. The comprehensive simulation results demonstrate that the proposed scheme is very efficient and can achieve a high message delivery ratio. We believe it can be used in many practical applications.

In wireless sensor networks (WSNs), providing source-location privacy through secure routing is one of the most prosperous techniques. In this paper, we propose a routing technique to provide adequate source-location privacy with low energy consumption. We introduce this technique as the Sink Toroidal Region (STaR) routing. With this technique, the source node randomly selects an intermediate node within a designed STaR area located around the SINK node. The STaR area is large enough to make it unpractical for an adversary to monitor the entire region. Furthermore, this routing protocol ensures that the intermediate node is neither too close, nor too far from the SINK node in relations to the entire network. While ensuring source location privacy, our simulation results show that the proposed scheme is very efficient and can be used for practical applications.

Wireless sensor networks (WSNs) have the potential to be widely used in many areas for unattended event monitoring. Mainly due to lack of a protected physical boundary, wireless communications are vulnerable to unauthorized interception and detection. Privacy is becoming one of the major issues that jeopardize the successful deployment of wireless sensor networks. While confidentiality of the message can be ensured through content encryption, it is much more difficult to adequately address the source-location privacy. For WSNs, source-location privacy service is further complicated by the fact that the sensor nodes consist of low-cost and low-power radio devices, computationally intensive cryptographic algorithms and large scale broadcasting-based protocols are not suitable for WSNs. In this paper, we propose source-location privacy schemes through routing to randomly selected intermediate node(s) before the message is transmitted to the SINK node. We first describe routing through a single a single randomly selected intermediate node away from the source node. Our analysis shows that this scheme can provide great local source-location privacy. We also present routing through multiple randomly selected intermediate nodes based on angle and quadrant to further improve the global source location privacy. While providing source-location privacy for WSNs, our simulation results also demonstrate that the proposed schemes are very efficient in energy consumption, and have very low transmission latency and high message delivery ratio. Our protocols can be used for many practical applications.

Wireless sensor networks (WSN) have the potential to be widely used in many areas for unattended event monitoring. Mainly due to lack of a protected physical boundary, wireless communications are vulnerable to unauthorized interception and detection. Privacy is becoming one of the major issues that jeopardize the successful deployment of wireless sensor networks. While confidentiality of the message can be ensured through content encryption, it is much more difficult to adequately address the source-location privacy. For WSN, source-location privacy service is further complicated by the fact that the sensor nodes consist of low-cost and low-power radio devices, computationally intensive cryptographic algorithms (such as public-key cryptosystems) and large scale broadcasting-based protocols are not suitable for WSN. In this paper, we propose a scheme to provide both content confidentiality and source-location privacy through routing to a randomly selected intermediate node (RRIN). While being able to provide source-location privacy for WSN, our simulation results also demonstrate that the proposed scheme is very efficient and can be used for practical applications.

Wireless sensor networks consist of low-cost and low- power radio devices and are deployed in open and unprotected areas. Privacy is becoming one of the major issues that jeopardize the successful deployment of wireless sensor networks. While confidentiality of the message can be ensured through content encryption, it is much more difficult to adequately address the source-location privacy. For wireless sensor networks, computationally intensive cryptographic algorithms (such as public-key cryptosystems) and large scale broadcasting-based protocols are not suitable. In this paper, we propose a scheme to provide source-location privacy through a three-phase routing: routing to a randomly selected intermediate node, routing in a network mix ring, and message forwarding to the SINK node. While being able to provide source-location privacy for WSN, our simulation results also demonstrate that the proposed scheme is very efficient and can be used for practical applications.

Communication privacy is becoming an essential security requirement for mission critical communications and communication infrastructure protection. This is especially true for mobile ad hoc networks (MANETs) due to mobility of the communication nodes and the nature of wireless communications. Existing research in privacy-preserving communications can largely be divided into two categories: cryptosystem-based techniques and broadcasting-based techniques. The cryptosystem-based techniques include mix-based systems and secure multiparty computation-based systems, originating from mixnet and DC-net respectively. All mix-based approaches require a trusted third party to provide the mix and are not quite feasible in MANET. However, DC-net based approaches suffer from transmission collision problem that cannot be easily resolved practically. Broadcasting based schemes provide communication privacy by mixing the real messages with dummy packets so that it is infeasible for the adversaries to identify the real packets and track the message source. However, the transmission of dummy messages not only increases the energy consumption significantly, but also increases the network collisions and decreases the packet delivery ratio. In this paper, we first propose a novel unconditionally secure source anonymous message authentication scheme (SAMAS) that enables messages to be released without relying on any trusted third parties. While providing source privacy, the proposed scheme can also provide message content authenticity. We then propose a novel communication protocol for MANET that can ensure communication privacy of both communication parties and their end-to-end routing. The proposed protocol can be used for critical infrastructure protection and secure file sharing. The security analysis demonstrates that the proposed protocol is secure against various attacks. The theoretical analysis and simulation show that the proposed scheme is efficient and can ensure high message delivery ratio.

Wireless sensor networks (WSN) have the potential to be widely used in many areas for unattended event monitoring. Mainly due to lack of a protected physical boundary, wireless communications are vulnerable to unauthorized interception and detection. Privacy is becoming one of the major issues that jeopardize the successful deployment of wireless sensor networks. While confidentiality of the message can be ensured through content encryption, it is much more difficult to adequately address the source-location privacy. For WSN, source-location privacy service is further complicated by the fact that the sensor nodes consist of low-cost and low-power radio devices, computationally intensive cryptographic algorithms (such as public-key cryptosystems) and large scale broadcasting-based protocols are not suitable for WSN. In this paper, we propose a scheme to provide both content confidentiality and source-location privacy through routing to a randomly selected intermediate node (RRIN) and a network mixing ring (NMR), where the RRIN provides local source- location privacy and NMR yields network-level (global) source- location privacy. While being able to provide source-location privacy for WSN, our simulation results also demonstrate that the proposed scheme is very efficient and can be used for practical applications.