arXiv subjects

Mehdi Bahrami

Publications and source records attributed to Mehdi Bahrami.

8 recordsLinked to original sources

arXiv2026

Kozuchi Agent: A Language-Agnostic Open-Weight Agent for Software Repair

Industrial software-engineering teams increasingly need LLM agents that turn bug reports into correct patches, yet benchmark-scale operation adds long horizons, tool-use discipline, context persistence, heterogeneous clusters, and evaluation reuse. We present Kozuchi Agent, a language-agnostic open-weight repair agent and CI-operated evaluation pipeline. Explicit phases, persistent state, deterministic tools, a model-independent action interface, and cross-agent test-time selection make runs auditable and repeatable. With locally hosted Qwen3.5-27B, no fine-tuning, and TTS@8, Kozuchi resolves 374/500 SWE-bench Verified instances on the official evaluator. Unchanged on Multi-SWE-bench Java, the same 27-billion-parameter agent resolves 41/128 instances (32.03%), ranking first among strict open-weight submissions and fourth of 42 overall; on Python it ranks 12th of 135 and first among open-weight systems. Per-phase behavior remains within +/-5 percentage points across languages. Remaining failures mainly reflect semantic correctness, Java-specific harness issues, and selection errors. Across both tracks, results compare favorably with open/local peers by parameter count. Analysis of candidate diversity, selector regret, and patch reliability shows that the remaining gap is primarily semantic correctness and selection rather than edit formatting or proprietary-model access. Operationally, reusable CI stages reduce operator touch-points from five to one across heterogeneous internal clusters.

cs.SE↗

arXiv2021

AugmentedCode: Examining the Effects of Natural Language Resources in Code Retrieval Models

Code retrieval is allowing software engineers to search codes through a natural language query, which relies on both natural language processing and software engineering techniques. There have been several attempts on code retrieval from searching snippet codes to function codes. In this paper, we introduce Augmented Code (AugmentedCode) retrieval which takes advantage of existing information within the code and constructs augmented programming language to improve the code retrieval models' performance. We curated a large corpus of Python and showcased the the framework and the results of augmented programming language which outperforms on CodeSearchNet and CodeBERT with a Mean Reciprocal Rank (MRR) of 0.73 and 0.96, respectively. The outperformed fine-tuned augmented code retrieval model is published in HuggingFace at https://huggingface.co/Fujitsu/AugCode and a demonstration video is available at: https://youtu.be/mnZrUTANjGs .

cs.SE↗

arXiv2021

PyTorrent: A Python Library Corpus for Large-scale Language Models

A large scale collection of both semantic and natural language resources is essential to leverage active Software Engineering research areas such as code reuse and code comprehensibility. Existing machine learning models ingest data from Open Source repositories (like GitHub projects) and forum discussions (like Stackoverflow.com), whereas, in this showcase, we took a step backward to orchestrate a corpus titled PyTorrent that contains 218,814 Python package libraries from PyPI and Anaconda environment. This is because earlier studies have shown that much of the code is redundant and Python packages from these environments are better in quality and are well-documented. PyTorrent enables users (such as data scientists, students, etc.) to build off the shelf machine learning models directly without spending months of effort on large infrastructure. The dataset, schema and a pretrained language model is available at: https://github.com/fla-sil/PyTorrent

cs.SE↗

arXiv2021

A Systematic Investigation of KB-Text Embedding Alignment at Scale

Knowledge bases (KBs) and text often contain complementary knowledge: KBs store structured knowledge that can support long range reasoning, while text stores more comprehensive and timely knowledge in an unstructured way. Separately embedding the individual knowledge sources into vector spaces has demonstrated tremendous successes in encoding the respective knowledge, but how to jointly embed and reason with both knowledge sources to fully leverage the complementary information is still largely an open problem. We conduct a large-scale, systematic investigation of aligning KB and text embeddings for joint reasoning. We set up a novel evaluation framework with two evaluation tasks, few-shot link prediction and analogical reasoning, and evaluate an array of KB-text embedding alignment methods. We also demonstrate how such alignment can infuse textual information into KB embeddings for more accurate link prediction on emerging entities and events, using COVID-19 as a case study.

cs.CL↗

arXiv2014

An overview to Software Architecture in Intrusion Detection System

Today by growing network systems, security is a key feature of each network infrastructure. Network Intrusion Detection Systems (IDS) provide defense model for all security threats which are harmful to any network. The IDS could detect and block attack-related network traffic. The network control is a complex model. Implementation of an IDS could make delay in the network. Several software-based network intrusion detection systems are developed. However, the model has a problem with high speed traffic. This paper reviews of many type of software architecture in intrusion detection systems and describes the design and implementation of a high-performance network intrusion detection system that combines the use of software-based network intrusion detection sensors and a network processor board. The network processor which is a hardware-based model could acts as a customized load balancing splitter. This model cooperates with a set of modified content-based network intrusion detection sensors rather than IDS in processing network traffic and controls the high-speed.

cs.SE↗

arXiv2014

A Novel Self-Recognition Method for Autonomic Grid Networks Case Study: Advisor Labor Law Software Application

Recently, Grid Computing Systems have provided wide integrated use of resources. Grid computing systems provide the ability to share, select and aggregate distributed resources as computers, storage systems or other devices in an integrated way. Grid computing systems have solved many problems in science, engineering and commerce fields. In this paper we introduce a self-recognition algorithm for grid network and introduced this algorithm to have exclusive management control on the autonomic grid networks. This algorithm is base on binomial heap to allocate and recognition any node in the grid. We try to using this algorithm in advisor labor law software application as case study and shown in this application how to use this method for any advisor application on the network. By this implementation model shown this method can get better answer to any question as a best labor law advisor.

cs.DC↗

arXiv2013

Cloud Template, a Big Data Solution

Today cloud computing has become as a new concept for hosting and delivering different services over the Internet for big data solutions. Cloud computing is attractive to different business owners of both small and enterprise as it eliminates the requirement for users to plan ahead for provisioning, and allows enterprises to start from the small and increase resources only when there is a rise in service demand. Despite the fact that cloud computing offers huge opportunities to the IT industry, the development of cloud computing technology is currently has several issues. This study presents an idea for introducing cloud templates which will be used for analyzing, designing, developing and implementing cloud computing systems. We will present a template based design for cloud computing systems, highlighting its key concepts, architectural principles and state of the art implementation, as well as research challenges and future work requirements. The aim of this idea is to provide a better understanding of the design challenges of cloud computing and identify important research directions in this big data increasingly important area. We will describe a series of studies by which we and other researchers have assessed the effectiveness of these techniques in practical situations. Finally, in this study we will show how this idea could be implemented in a practical and useful way in industry.

cs.DC↗

arXiv2008

A New Method for Knowledge Representation in Expert System's (XMLKR)

Knowledge representation it is an essential section of a Expert Systems, Because in this section we have a framework to establish an expert system then we can modeling and use by this to design an expert system. Many method it is exist for knowledge representation but each method have problems, in this paper we introduce a new method of object oriented by XML language as XMLKR to knowledge representation, and we want to discuss advantage and disadvantage of this method.

cs.DC↗