Mohammad Izadi

Mohammad Reza Besharati, Mohammad Izadi, Ehsaneddin Asgari · berkeley

Underlying computational model has an important role in any computation. The state and transition (such as in automata) and rule and value (such as in Lisp and logic programming) are two comparable and counterpart computational models. Both of deductive and model checking verification techniques are relying on a notion of state and as a result, their underlying computational models are state dependent. Some verification problems (such as compliance checking by which an under compliance system is verified against some regulations and rules) have not a strong notion of state nor transition. Behalf of it, these systems have a strong notion of value symbols and declarative rules defined on them. SARV (Stateless And Rule-Based Verification) is a verification framework that designed to simplify the overall process of verification for stateless and rule-based verification problems (e.g. compliance checking). In this paper, a formal logic-based framework for creating intelligent compliance checking systems is presented. We define and introduce this framework, report a case study and present results of an experiment on it. The case study is about protocol compliance checking for smart cities. Using this solution, a Rescue Scenario use case and its compliance checking are sketched and modeled. An automation engine for and a compliance solution with SARV are introduced. Based on 300 data experiments, the SARV-based compliance solution outperforms famous machine learning methods on a 3125-records software quality dataset.

Mohammad Izadi, Mehran Safayani, Abdolreza Mirzaei

Efficient real-time traffic prediction is crucial for reducing transportation time. To predict traffic conditions, we employ a spatio-temporal graph neural network (ST-GNN) to model our real-time traffic data as temporal graphs. Despite its capabilities, it often encounters challenges in delivering efficient real-time predictions for real-world traffic data. Recognizing the significance of timely prediction due to the dynamic nature of real-time data, we employ knowledge distillation (KD) as a solution to enhance the execution time of ST-GNNs for traffic prediction. In this paper, We introduce a cost function designed to train a network with fewer parameters (the student) using distilled data from a complex network (the teacher) while maintaining its accuracy close to that of the teacher. We use knowledge distillation, incorporating spatial-temporal correlations from the teacher network to enable the student to learn the complex patterns perceived by the teacher. However, a challenge arises in determining the student network architecture rather than considering it inadvertently. To address this challenge, we propose an algorithm that utilizes the cost function to calculate pruning scores, addressing small network architecture search issues, and jointly fine-tunes the network resulting from each pruning stage using KD. Ultimately, we evaluate our proposed ideas on two real-world datasets, PeMSD7 and PeMSD8. The results indicate that our method can maintain the student's accuracy close to that of the teacher, even with the retention of only 3% of network parameters.

Hosein Hasani, Amirmohammad Izadi, Fatemeh Askari et al.

This paper examines how large language models (LLMs) and large vision-language models (LVLMs) represent and compute numerical information in counting tasks. We use controlled experiments with repeated textual and visual items and analyze model behavior through causal mediation and activation patching. To this end, we design a specialized tool, CountScope, for mechanistic interpretability of numerical content. Results show that individual tokens or visual features encode latent positional count information that can be extracted and transferred across contexts. Layerwise analyses reveal a progressive emergence of numerical representations, with lower layers encoding small counts and higher layers representing larger ones. We identify an internal counter mechanism that updates with each item, stored mainly in the final token or region and transferable between contexts. In LVLMs, numerical information also appears in visual embeddings, shifting between background and foreground regions depending on spatial composition. Models rely on structural cues such as separators in text, which act as shortcuts for tracking item counts and influence the accuracy of numerical predictions. Overall, counting emerges as a structured, layerwise process in LLMs and follows the same general pattern in LVLMs, shaped by the properties of the vision encoder.

Hosein Hasani, Amirmohammad Izadi, Fatemeh Askari et al.

Large vision-language models (LVLMs) show strong performance across multimodal benchmarks but remain limited in structured reasoning and precise grounding. Recent work has demonstrated that adding simple visual structures, such as partitions and annotations, improves accuracy, yet the internal mechanisms underlying these gains remain unclear. We investigate this phenomenon and propose the concept of Grounding IDs, latent identifiers induced by external cues that bind objects to their designated partitions across modalities. Through representation analysis, we find that these identifiers emerge as robust within-partition alignment in embedding space and reduce the modality gap between image and text. Causal interventions further confirm that these identifiers mediate binding between objects and symbolic cues. We show that Grounding IDs strengthen attention between related components, which in turn improves cross-modal grounding and reduces hallucinations. Taken together, our results identify Grounding IDs as a key symbolic mechanism explaining how external cues enhance multimodal binding, offering both interpretability and practical improvements in robustness.

Mohammad Izadi, Mehran Safayani

Autism Spectrum Disorder (ASD) is a complex neurodevelopmental condition marked by disruptions in brain connectivity. Functional MRI (fMRI) offers a non-invasive window into large-scale neural dynamics by measuring blood-oxygen-level-dependent (BOLD) signals across the brain. These signals can be modeled as interactions among Regions of Interest (ROIs), which are grouped into functional communities based on their underlying roles in brain function. Emerging evidence suggests that connectivity patterns within and between these communities are particularly sensitive to ASD-related alterations. Effectively capturing these patterns and identifying interactions that deviate from typical development is essential for improving ASD diagnosis and enabling biomarker discovery. In this work, we introduce ASDFormer, a Transformer-based architecture that incorporates a Mixture of Pooling-Classifier Experts (MoE) to capture neural signatures associated with ASD. By integrating multiple specialized expert branches with attention mechanisms, ASDFormer adaptively emphasizes different brain regions and connectivity patterns relevant to autism. This enables both improved classification performance and more interpretable identification of disorder-related biomarkers. Applied to the ABIDE dataset, ASDFormer achieves state-of-the-art diagnostic accuracy and reveals robust insights into functional connectivity disruptions linked to ASD, highlighting its potential as a tool for biomarker discovery.

Mohammad Reza Besharati, Mohammad Izadi

Video games, just like any other media have both explicit and implicit messages, and they can have impact on physical and mental health of the users. These impacts can be positive or negative. The impacts, the implications and the meanings which exist in a game can be very widespread, multilayered and complicated. To investigate and guarantee the health of these video games, it is necessary to be able to estimate, assess and determine the implications of video games (from different perspectives). A common approach for studying complicated and multilayered phenomenon is formal semantics. Formal and rigorous methods can help in assessment and supplying the health of video games. In this article, an organizing for this assessment is proposed which is based on formal and rigorous methods and it considers various beneficiaries concerns. Moreover, a technological solution is presented which is based on system compliance to meanings, model checking methods and logical solution. The proposed organizing has several features such as: agility, flexibility, scalability, repeatability of reviews, transparency, adaptation, available details for reviews, assessing various layers and implicit and explicit implications of system of the game, avoiding subjectivity or individual skills, relying on rules and regulations, ability to plan for beneficiaries because of its transparency and employment for specialists.

Nafiseh Jafari, Mohammad Reza Besharati, Mohammad Izadi et al.

One of the pillars of any machine learning model is its concepts. Using software engineering, we can engineer these concepts and then develop and expand them. In this article, we present a SELM framework for Software Engineering of machine Learning Models. We then evaluate this framework through a case study. Using the SELM framework, we can improve a machine learning process efficiency and provide more accuracy in learning with less processing hardware resources and a smaller training dataset. This issue highlights the importance of an interdisciplinary approach to machine learning. Therefore, in this article, we have provided interdisciplinary teams' proposals for machine learning.

Mohammad Reza Besharati, Mohammad Izadi

Reo is a formal coordination language. In order to assess and evaluate its capabilities, we need a multi-perspective Language Evaluation Framework. Langar (Language Analysis for Reo) is a framework aimed to provide such an evaluation method. In this paper, we introduce Langar. Based on a review on various language evaluation methods, a tool-kit for useful evaluation techniques are provided. After Reo Evaluation, this method and tool-kit also could be used for another programming, computational and even natural languages. Furthermore, two suggestions for some future efforts and directions are provided for software engineering and software methodology communities.

Vahid Shahrivari, Mohammad Mahdi Darabi, Mohammad Izadi

The Internet has become an indispensable part of our life, However, It also has provided opportunities to anonymously perform malicious activities like Phishing. Phishers try to deceive their victims by social engineering or creating mock-up websites to steal information such as account ID, username, password from individuals and organizations. Although many methods have been proposed to detect phishing websites, Phishers have evolved their methods to escape from these detection methods. One of the most successful methods for detecting these malicious activities is Machine Learning. This is because most Phishing attacks have some common characteristics which can be identified by machine learning methods. In this paper, we compared the results of multiple machine learning methods for predicting phishing websites.

Mohammad Reza Besharati, Mohammad Izadi

Instead of proofs or logical evaluations, compliance assessment could be done by benchmarking. Benchmarks, in their nature, are applied. So a set of benchmarks could shape an applied solution for compliance assessment. In this paper, we introduce the KARB solution: Keeping away compliance Anomalies by Rule-based Benchmarking. By rule-based benchmarking, we mean evaluation of under-compliance-system by its symbolic specification and by using a set of symbolic rules (on behalf of semantic logic of evaluation). In order to demonstrate and investigate the manner of KARB solution, we conducted a case study. The IR-QUMA study (Iranian Survey on Quality in Messenger Apps) is defined to evaluate the quality of some messenger apps. the results of evaluations suggest that the Hybrid Method of DD-KARB (with combination of semantics-awareness and data-drivenness) is more effective than solo-methods and could compute a somehow good estimation for messenger-apps user quality scores. So DD-KARB could be considered as a method for quality benchmarking in this technical context.

Hossein Keshavarz, Shohreh Tabatabayi Seifi, Mohammad Izadi

In this paper, we propose a novel approach for measuring the degree of similarity between categories of two pieces of Persian text, which were published as descriptions of two separate advertisements. We built an appropriate dataset for this work using a dataset which consists of advertisements posted on an e-commerce website. We generated a significant number of paired texts from this dataset and assigned each pair a score from 0 to 3, which demonstrates the degree of similarity between the domains of the pair. In this work, we represent words with word embedding vectors derived from word2vec. Then deep neural network models are used to represent texts. Eventually, we employ concatenation of absolute difference and bit-wise multiplication and a fully-connected neural network to produce a probability distribution vector for the score of the pairs. Through a supervised learning approach, we trained our model on a GPU, and our best model achieved an F1 score of 0.9865.

MohammadReza Besharati, Mohammad Izadi

Measuring text complexity is an essential task in several fields and applications (such as NLP, semantic web, smart education, etc.). The semantic layer of text is more tacit than its syntactic structure and, as a result, calculation of semantic complexity is more difficult than syntactic complexity. While there are famous and powerful academic and commercial syntactic complexity measures, the problem of measuring semantic complexity is still a challenging one. In this paper, we introduce the DAST model, which stands for Deciding About Semantic Complexity of a Text. DAST proposes an intuitionistic approach to semantics that lets us have a well-defined model for the semantics of a text and its complexity: semantic is considered as a lattice of intuitions and, as a result, semantic complexity is defined as the result of a calculation on this lattice. A set theoretic formal definition of semantic complexity, as a 6-tuple formal system, is provided. By using this formal system, a method for measuring semantic complexity is presented. The evaluation of the proposed approach is done by a set of three human-judgment experiments. The results show that DAST model is capable of deciding about semantic complexity of text. Furthermore, the analysis of the results leads us to introduce a Markovian model for the process of common-sense, multiple-steps and semantic-complexity reasoning in people. The results of Experiments demonstrate that our method outperforms the random baseline with improvement in better precision and competes with other methods by less error percentage.