Existing code generation benchmarks, such as HumanEval, primarily evaluate universal algorithmic problems that do not address region-localized requirements. In realworld software development, localized knowledge, such as legal systems, cultural practices, and local language-specific processing is often required. To address this gap, we propose SakuraEval, the first region-localized code generation benchmark designed to evaluate Large Language Models, (LLMs,) performance in generating code that requires Japanese-localized knowledge. As a demonstration, of its effectiveness, we evaluated five LLMs with different language backgrounds using SakuraEval alongside HumanEval and JHumanEval. The results showed that all models exhibited significantly different tendencies in SakuraEval compared to existing benchmarks, with notable changes in their rankings that highlight each model’s Japanese-localized capabilities. SakuraEval provides valuable insights into the development of region-aware LLMs and underscores the importance of localized knowledge in code generation tasks.

ChatGPT presents emerging opportunities in software development, yet its capabilities for understanding code remain largely understudied. This study aims to focus on the logical aspect of code comprehension of ChatGPT by examining its performance in detecting and fixing bugs. Our preliminary results suggest that ChatGPT seems to correct code in a different way than human logical steps.

Pre-trained language models have achieved many prominent results in natural language processing. Since software engineering widely includes many natural language documents, the application of pre-trained language models have received much attention in software engineering tasks. However, pretraining a large volume of source code requires a huge amount of computational resources and time. In this study, we propose an additional pre-training approach to a well-trained language model. Our initial results on mT5, multilingual T5 with an additional pretraining of Python code shows improved performance on multiple software engineering tasks including code generation, code summarisation, code repair, and error diagnosis.

Parsing expression grammars (PEGs) are a powerful and popular foundation for describing syntax. Despite PEGs' expressiveness, they cannot recognize many syntax patterns of popular programming languages. Typical examples include typedef-defined names in C/C++ and here documents appearing in many scripting languages. We use a single unified state representation, called a symbol table, to capture various context-sensitive patterns. Over the symbol table, we design a small set of restricted semantic predicates and actions. The extended PEGs are called SPEGs, and are designed to be safe in contexts of backtracking and the linear time guarantee of packrat parsing. This paper will show that SPEGs have improved the expressive power in such ways that they recognize practical context-sensitive grammars, including back referencing, indentation-based code layout, and contextual keywords.

Modern computer systems are increasing in complexity, spread, and scale in order to meet the diverse and sophisticated needs of the users. In the development of these systems, we inevitably use legacy codes and off-the shelf software as black box software in order to shorten the development time and lower the development cost. These systems are often connected via a network to utilize services provided by other systems, whereas services and network performance may change while in operation. We often need to change the specification and implementation of a system due to the changes of environments and users' requirements. In addition, threats caused by viruses and unauthorized accesses have to be properly removed. Therefore, modern computer systems inherently involve incompleteness of specifications and implementations and uncertainty of environments and requirements.

In this paper we propose an online lightweight anomaly symptom detection and process's resource usage control mechanism. Our system collects fine-grain resource information that can reflect the subtle changes of the application's behavior. Then it creates models with a learning based algorithm without manual configurations. If an anomaly symptom is detected, the automatic procedure will start. The system will control the suspected application's resource use by limiting the upper bound resource of the process. The method will make the application yield its CPU to the administrative inspection. In this paper, we described whole architecture of the system and evaluate it with the non deterministic and deterministic failure. Our experimental results indicate that our prototype system is able to detect non deterministic failure with high precision in anomaly training and control it's resource use with an overhead of about 1%.

A key challenge in ubiquitous computing is how to manage ambient information about the real-world environment where a variety of devices appear and disappear. Traditional capability description systems are inadequate in capturing dynamic changes in the environment. We propose the Toibox DVMS system, a multi-embedded database system that provides an integrated database view of computing devices. The integrated views are materialized; updates over the view provoke relevant operations for the specified devices. The materialized views are decentralized and incrementally maintained by data dissemination. This work presents the design of the Toibox system, based on our initial prototype implementations.

The expanding market for digital content increases the demand to transfer electronic value through open networks. This paper describes the Ticket Transfer Protocol (TTP), which enables electronically transferable tickets to directly move between two personal tamper-proof devices. The applied two-phase commit protocol for the transfer realizes the ACID (atomicity, consistency, isolation, durability) transaction properties without any external coordinators. Moreover, this paper addresses the initial implementation of a secure TTP, enabling the detection of message replays by attackers, and evaluates the size of its logging field in the transaction. We conclude that TTP has enough reliability to transfer electronic tickets directly between personal ticket wallets.

Internet commerce is increasing the demands of service integrations by sharing XML-based catalogs. We propose the PCO (Portable Compound Object) data model supporting semantic inheritance to ensure the synonymy of heterogeneous semantics among distributed schemas that different authors define independently. Also, the PCO model makes semantic relationships independent of an initial class hierarchy, and it enables rapid schema evolution across the entire Internet business. This preserves semantic interoperability without changing pre-defined classes. We have also encoded the PCO model into two XML-based languages: the PCO Specification Language (PSL) and the Portable Composite Language (PCL). This paper demonstrates that intermediaries defining service semantics in PSL can automatically integrate multiple suppliers' PCL catalogs for their agent-mediated services.

Proposes a new decentralized architecture for electronic markets, called Digiket (DIGItal marKET), as an alternative to traditional virtual malls. The present Internet commerce consists of relatively small closed markets; so, new frameworks that enable Web-based stores and malls to share their services worldwide are strongly desired. Digiket separates electronic commerce functions (such as content authoring, catalog delivery, intermediary services, shopping assistants and transaction processing) from traditional mall sites, and decentralizes each of them independently. Digiket's architecture is unique in that it requires no central coordinators for the distributed components. Instead, the PCO (Portable Compound Object) language, a lingua franca for Digiket, connects the components to each other. We have also implemented the PCO language and a user agent client named OpenBasket, to demonstrate a shopping application on Digiket. The demonstration shows that a consumer can order arbitrary goods described in PCO without vendor barriers.