Skip to content
Open access

TensorLock: Recovering Model Dependency for Model Supply Chain

Oct 2026 · Proceedings of the ACM on Software Engineering · Vol 3, pp. 641 - 663 · 0 citations · 61 references

TL;DR

This work proposes a novel model dependency recovering framework Tensorlock, which substantially outperforms state-of-the-art approaches in clustering and dependency identification, and applies it to 289 supposedly isolated models and recover 189 previously missing model dependencies.

Abstract

The public models on the model hosting platforms have undergone exponential growth, allowing developers to build upon existing models rather than training from scratch. These models are continuously reused, modified, and re-distributed similar to traditional software components, breeding a dense and rapidly evolving model supply chain. However, while enjoying the benefits of model reuse, developers also inherit supply chain risks ranging from legal liabilities to security threats. To mitigate these risks, a well-established model dependency graph can significantly benefit supply chain risk governance. Unfortunately, although model hosting platforms offer mechanisms for dependency disclosure, such declarations are optional and frequently missing. To address this challenge, we propose a novel model dependency recovering framework Tensorlock. It works in two phases; i.e., (1) model clustering, and (2) type-aware dependency identification within these clusters. In the first phase, Tensorlock performs connectivity-based clustering to accommodate the open-ended dependency topology, grouping models with dependency relations. In the second phase, Tensorlock employs a divide-and-conquer strategy, leveraging distinct type-specific fingerprints to first identify data-free dependencies (Quantization and Merging), and then resolve data-driven Fine-Tuning dependencies. Our evaluation demonstrates that Tensorlock substantially outperforms state-of-the-art approaches, achieving an ARI of 0.96 in clustering and a DF1 of 0.82 in dependency identification, improving over the best baselines by at least 39% and 193%, respectively. Additionally, we apply Tensorlock to 289 supposedly isolated models and recover 189 previously missing model dependencies, with 42 model authors confirming our findings.

Read PDF

Similar papers

#artificial intelligence Preprint Oct 2026

Dual-Scale Relational Graph Transformers for Ecosystem-Aware Fraud Detection

Account takeover (ATO) fraud is a growing threat to digital banking, requiring effective detection while minimizing friction for legitimate customers. Production systems predominantly rely on tabular models that score sessions in isolation, discarding the relational structure of the underlying interaction network. Alth...

Mohsen Nayebi Kerdabadi, Xin-Rou Li, Yao Xiao et al. · 0 citations
Book Open access Oct 2026

S2CAgent: A Self-Learning Agent Knowledge Graph Framework for Software Supply Chain Analysis

The extensive reuse of open-source software (OSS) accelerates development. However, it also creates increasingly complex software supply chains (SSCs) and intensifies the attacker-defender asymmetry, making attacks "low-cost, high-impact, and hard-to-detect." Software Bill of Materials (SBOMs) and Software Composition...

You-Fei Huang, Zhong Li, Min-Xue Pan et al. · 0 citations
#large language models Book Open access Oct 2026

Characterizing the Ecosystem of Open-Source Uncensored Large Language Models: Supply Chain, Governance, and Risks

Open-source uncensored large language models (ULLMs) are increasingly available in public LLM ecosystems. By removing or weakening safety guardrails, these models can comply with harmful requests, generate unsafe content, and be readily reused in downstream systems. While prior work has demonstrated the misuse potentia...

Zilong Lin, Zi-Chuan Li, Xiao-Jing Liao et al. · 0 citations
Open access Oct 2026

Dynamic Graph Learning for Core Data Asset Discovery Using Data Lineage Graphs

The rapid growth of the digital economy has made the accurate identification of Core Data Assets (CDAs) increasingly important for enterprise data governance. Existing approaches, however, often rely on static, homogeneous dependency modeling or node-centric analysis, limiting their ability to capture complex value pro...

Yi Pan, Yu-Fei Lei, R. Li · 0 citations
Book Open access Nov 2026

Can Frontier LLMs Capture $40 Billion in Supply Chain Value?

Large Language Models (LLMs) perform well on many NLP tasks, but their ability to reason about complex supply chain relationships remains unclear. We introduce the Supply Chain Intuition Exam (SCIE), a benchmark for evaluating whether LLMs can identify and infer direct and indirect supplier relationships. We compare st...

Chung-Chi Chen, Ichiro Kobayashi, Yusuke Miyao et al. · 0 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.