This paper researches on the power consumption of AIGC applications based on LLM with different parameter scales across different hardware platforms. Artificial Intelligence Generated Content (AIGC) represents a leading-edge application of AI technology, primarily driven by large language models (LLMs) and their associated technologies. The deployment of LLM typically relies on critical facilities with three layers, i.e., the hardware, model, and application layers. This empirical study aims to identify key factors in power consumption when a large model is serving in the inference stage, which will hint the insights for improving the energy efficiency of computational infrastructures. In the context of the "dual carbon" goals, i.e., carbon peaking and carbon neutrality, this study aims to find an effective way to reduce the energy cost of AIGC applications, thereby supporting sustainable AI development in industry.

Blockchain has been extensively used on the Internet of Things (IoT). However, several problems exist in current blockchain systems that limit their usability in IoT networks. First, Proof-of-Work (PoW) consensus most used in current blockchain systems generates excessive energy consumption. Second, the single digital token system limits the development of blockchain in multiple IoT applications. Finally, the large amount of data generated by IoT devices and stored in blockchains has huge storage requirements. To deal with these problems, we propose HuaBaseChain—an extensible blockchain with high performance. HuaBaseChain has three features. First, HuaBaseChain uses a novel Proof-of-Participation (PoP) consensus to reduce energy consumption. Second, HuaBaseChain extends the single digital token model to a multidimensional digital token model, enabling more varied applications that can be tailored to the needs of IoT device managers. Third, HuaBaseChain devises an underlying Merkle forest data structure to reduce storage requirements. HuaBaseChain is suited for IoT applications that contain a variety of IoT devices and generate massive amounts of data. We demonstrate the efficiency of HuaBaseChain on IoT data sets and the significantly larger Bitcoin data set. Our experiments show that HuaBaseChain runs efficiently, with significantly reduced energy consumption, greater efficiency, lower storage requirements, and higher query speed.

The healthcare industry faces serious problems with health data. Firstly, health data is fragmented and its quality needs to be improved. Data fragmentation means that it is difficult to integrate the patient data stored by multiple health service providers. The quality of these heterogeneous data also needs to be improved for better utilization. Secondly, data sharing among patients, healthcare service providers and medical researchers is inadequate. Thirdly, while sharing health data, patients' right to privacy must be protected, and patients should have authority over who can access their data. In traditional health data sharing system, because of centralized management, data can easily be stolen, manipulated. These systems also ignore patient's authority and privacy. Researchers have proposed some blockchain-based health data sharing solutions where blockchain is used for consensus management. Blockchain enables multiple parties who do not fully trust each other to exchange their data. However, the practice of smart contracts supporting these solutions has not been studied in detail. We propose CrowdMed-II, a health data management framework based on blockchain, which could address the above-mentioned problems of health data. We study the design of major smart contracts in our framework and propose two smart contract structures. We also introduce a novel search contract for searching patients in the framework. We evaluate their efficiency based on the execution costs on Ethereum. Our design improves on those previously proposed, lowering the computational costs of the framework. This allows the framework to operate at scale and is more feasible for widespread adoption.

The DaaS (Data as a Service) is an Internet-perception big data system based on AI, which is built by "Think Tank 2861 Project Team". This is an Internet-area, data-based, and neural feedback system for the Internet information in China. It takes Internet activities as the input, and processes through AI algorithms and machine learning framework to generate the output, based on which building the real-time macro economics and society big data for about 9.8 million grids in China and its intelligent applications. DaaS covers all 2,861 administrative districts and counties in the country and is refined to geography grid of one square kilometer granularity. The real-time objective information generated by distributed AI algorithms, that are constantly trained and calibrated, is of great value in scientific research and commercial applications.

Tourism has become the world's largest economy industry. More and more people share their travelogues on travel websites. Recommender system is an effective tool to provide travel services (e.g., Landscapes selection) for tourists. Many recommender systems are based on travel data that are supplied by travel agencies, and provide travel packages from a fixed package set, which bring two challenges for travel package recommender system. One is how to generate more travel packages. The other is how to measure more fine-grained user similarity. To address these challenges, we develop a package generation and recommendation framework to help travelers select landscapes. Firstly, we propose a Fuzzy Clustering based Package Generation algorithm (FCPG) to generate new travel packages to improve the overall recommendation effectiveness. Then, we develop a Dual Topic Model based Package Recommendation algorithm (DTMPR). It considers two user-related topics (travel seasons and areas), and provides more fine-grained user similarity measure. Experimental results show the superiority of our framework in comparison with the state-of-the-art methods.

Influence maximization modeling and analyzing is an important problem in Online Social Networks (OSNs). Influential nodes provide information on hot topics and forward interesting information, which can yield great influence on other nodes in OSNs. For word-of-mouth viral marketing, it is critical to find influential nodes within budgets. However, finding the optimal solution to maximize the influence spread in a given OSN has been proven to be NP-Hard. The existing algorithms suffer the following three defects: (1) need acquire the topological structure of the network, which is impractical for the continuously changing networks in real life, (2) are easy to get trapped in Rich Club phenomena, (3) can not balance very well between influence spread and running time. To solve this challenging problem, based on the randomly heuristic algorithm and the scale-free property of Complex Networks, we propose RMDN (Random Maximal Degree Neighbor) and its improved version RMDN++. We prove the feasibility and scalability of RMDNs in theory. Five real datasets are used to test the effectiveness and efficiency of RMDNs under two different influence diffusion models. The result shows that our methods have a comparable performance in terms of influence spread as state-of-the-art algorithms, but decrease the time by 1-2 orders of magnitude.

This article is composed in the background of the study of scientific field of coauthors phenomenon factual basis. By the study of massive amounts of relational data, it provides us with major significances theoretically and practically on retrieving and obtaining professionally academic information and getting knowing of academic development trend of miscellaneous fields. In process of studying this type of project, the problem of cluttering for coauthors that are in the data is involved. However, it is hard to meet the need of implementing the analysis of massive amounts of data cluttering by the existing cluttering software and algorithms, for this reason, finding an approach to deal with this kind of question is toughly important. To solve this question, this article presents an optimized Bisecting K-Means (BKM) clustering algorithm based on Hadoop and states the fashion of how to optimize the algorithm and the key point of implementing in details after analyzing the status quo related to this study. Estimating the complexity of the algorithm by experiments indicates the current problems and the direction for the future study.

With the development of cloud computing technologies, big data processing is becoming more and more important. How to mine and analyze massive data is facing a very big challenge. In this paper, we proposed an efficient massive data mining and analysis framework Data Cloud on large clusters. The most important part of Data Cloud is the Rabbit. It is a kind of massive data mining and analysis processing plan framework on the large clusters like the Pig and Hive. We make a detail analysis about the Rabbit plan.

With the development of IoT technologies, more and more medial sensors have been used to monitor people's health. In this paper, we design a kind of semantic medical monitoring system model in the cloud based on the IoT sensors. Lots of IoT sensors will accept the massive sensor data every time. All these massive sensor data will be stored in the HDFS and some information will be stored into the HUADING-S, a column-based database. All these HUADING-S data will connect to the medical rule engine. When the users' or patients' health indicators' data is beyond the normal range, the medical rule engine will send the alert information to the users or patients. We design two algorithms: (1) massive semantic medical rules processing algorithm without external communication and (2) massive semantic medical rules processing algorithm with external communication. Our simulation experiment shows that the algorithm 2 will reduce the time cost a lot and improve the executive efficiency.

In this paper, we design a kind of big data processing framework SemanMR (Semantic MapReduce). SemanMR is a programming framework based on the Hadoop MapReduce programming model. SemanMR provide a kind of bid data processing mechanism based on the metadata cluster of distributed file systems or cloud databases. In addition, we add some semantic index on the big data, and so it will improve our processing efficiency in SemanMR. SemanMR is a kind of big data processing internetware in the cloud environment.

It is now widely recognized that appropriate measures are required for digital preservation to ensure that digital data can be accessed and used currently and in the future. Among all the risks of digital preservation, format obsolescence is one of the most important. There have been several projects or initiatives dealing with the measurement method of format obsolescence risk, but there has been no mechanism to quantify the preservation risk or durability of digital formats based on a self-improving assessment model, executed with the aid of computers. This paper deals with a methodology for measuring the preservation durability of digital formats, especially for their risk assessment. This method is based on a quantitative assessment model for format risk, and can shift the non-quantifiable knowledge or experiences of field experts to a machine identifiable and processible form, or ‘risk scores’. Results can be recognized and communicated by computers automatically and formally, which can assist in the automatic/semi-automatic risk management for digital preservation, sharing this quantified knowledge among communities. Because technologies are changing quickly, the quantitative assessment model for risks will not be a status quo situation. Thus, also presented is a method to fine tune the quantitative assessment model for risk of formats through a self-learning and self-improving style.