arXiv subjects

Morteza Rakhshaninejad

Publications and source records attributed to Morteza Rakhshaninejad.

4 recordsLinked to original sources

arXiv2026

When should we trust the annotation? Selective prediction for molecular structure retrieval from mass spectra

Machine learning methods for identifying molecular structures from tandem mass spectra (MS/MS) have advanced rapidly, yet current approaches still exhibit significant error rates. In high-stakes applications such as clinical metabolomics and environmental screening, incorrect annotations can have serious consequences, making it essential to determine when a prediction can be trusted. In this work, we introduce a selective prediction framework for molecular structure retrieval from MS/MS spectra, separating low-risk predictions automatically from lower-confidence predictions. We formulate the problem within the risk-coverage tradeoff framework and systematically evaluate uncertainty quantification strategies at three levels: input-level distance in the learned representation, fingerprint-level uncertainty over predicted molecular fingerprint bits, and retrieval-level uncertainty over candidate rankings. We compare scoring functions including first-order confidence measures, aleatoric and epistemic uncertainty estimates from second-order distributions, as well as distance-based measures in the latent space. All experiments are conducted on the MassSpecGym benchmark. Our analysis reveals that while fingerprint-level uncertainty scores are poor proxies for retrieval success, retrieval-level total uncertainty provides the overall strongest rejection criterion, and first-order confidence measures are computationally inexpensive, strong baselines. We demonstrate that by applying distribution-free risk control via generalisation bounds, practitioners can specify a tolerable error rate and obtain a subset of annotations satisfying that constraint with high probability.

cs.LG↗

arXiv2025

A Language Model-Driven Semi-Supervised Ensemble Framework for Illicit Market Detection Across Deep/Dark Web and Social Platforms

Illegal marketplaces have increasingly shifted to concealed parts of the internet, including the deep and dark web, as well as platforms such as Telegram, Reddit, and Pastebin. These channels enable the anonymous trade of illicit goods including drugs, weapons, and stolen credentials. Detecting and categorizing such content remains challenging due to limited labeled data, the evolving nature of illicit language, and the structural heterogeneity of online sources. This paper presents a hierarchical classification framework that combines fine-tuned language models with a semi-supervised ensemble learning strategy to detect and classify illicit marketplace content across diverse platforms. We extract semantic representations using ModernBERT, a transformer model for long documents, finetuned on domain-specific data from deep and dark web pages, Telegram channels, Subreddits, and Pastebin pastes to capture specialized jargon and ambiguous linguistic patterns. In addition, we incorporate manually engineered features such as document structure, embedded patterns including Bitcoin addresses, emails, and IPs, and metadata, which complement language model embeddings. The classification pipeline operates in two stages. The first stage uses a semi-supervised ensemble of XGBoost, Random Forest, and SVM with entropy-based weighted voting to detect sales-related documents. The second stage further classifies these into drug, weapon, or credential sales. Experiments on three datasets, including our multi-source corpus, DUTA, and CoDA, show that our model outperforms several baselines, including BERT, ModernBERT, DarkBERT, ALBERT, Longformer, and BigBird. The model achieves an accuracy of 0.96489, an F1-score of 0.93467, and a TMCC of 0.95388, demonstrating strong generalization, robustness under limited supervision, and effectiveness in real-world illicit content detection.

cs.CL↗

arXiv2025

Conformal Prediction for Uncertainty Estimation in Drug-Target Interaction Prediction

Accurate drug-target interaction (DTI) prediction with machine learning models is essential for drug discovery. Such models should also provide a credible representation of their uncertainty, but applying classical marginal conformal prediction (CP) in DTI prediction often overlooks variability across drug and protein subgroups. In this work, we analyze three cluster-conditioned CP methods for DTI prediction, and compare them with marginal and group-conditioned CP. Clusterings are obtained via nonconformity scores, feature similarity, and nearest neighbors, respectively. Experiments on the KIBA dataset using four data-splitting strategies show that nonconformity-based clustering yields the tightest intervals and most reliable subgroup coverage, especially in random and fully unseen drug-protein splits. Group-conditioned CP works well when one entity is familiar, but residual-driven clustering provides robust uncertainty estimates even in sparse or novel scenarios. These results highlight the potential of cluster-based CP for improving DTI prediction under uncertainty.

cs.LG↗

arXiv2023

A Comprehensive Bibliometric Analysis on Social Network Anonymization: Current Approaches and Future Directions

In recent decades, social network anonymization has become a crucial research field due to its pivotal role in preserving users' privacy. However, the high diversity of approaches introduced in relevant studies poses a challenge to gaining a profound understanding of the field. In response to this, the current study presents an exhaustive and well-structured bibliometric analysis of the social network anonymization field. To begin our research, related studies from the period of 2007-2022 were collected from the Scopus Database then pre-processed. Following this, the VOSviewer was used to visualize the network of authors' keywords. Subsequently, extensive statistical and network analyses were performed to identify the most prominent keywords and trending topics. Additionally, the application of co-word analysis through SciMAT and the Alluvial diagram allowed us to explore the themes of social network anonymization and scrutinize their evolution over time. These analyses culminated in an innovative taxonomy of the existing approaches and anticipation of potential trends in this domain. To the best of our knowledge, this is the first bibliometric analysis in the social network anonymization field, which offers a deeper understanding of the current state and an insightful roadmap for future research in this domain.

cs.SI↗