Beining Yang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Beining Yang has a total of 20 co-authors. Here are the top co-authors by number of shared publications:
- Yang Cao
University of Edinburgh, Edinburgh, United Kingdom
2 shared publications - Chenhui Zhang
ZHIPU AI, Beijing, China
1 shared publication - Chunming Hu
Beihang University, Beijing, China
1 shared publication - Evgeny Kharlamov
Bosch Center for Artificial Intelligence, Robert Bosch GmbH, Renningen, Germany
1 shared publication - Hao Peng
Beihang University, Beijing, China
1 shared publication - Huanjing Zhao
Tsinghua University, Beijing, China
1 shared publication
Latest Publications
- Open Access
- 24 September 2026
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Management of Data
- Vol. 4 (4) , 1-26
Metric Algebra: Embedding-Independence in Vector Databases
Vector databases are increasingly central to retrieval applications, yet deployments remain embedding-siloed: different databases store vectors produced by different embedding models and ANN pipelines, making cross-database search ill-defined or forcing expensive re-embedding and index rebuilds. We propose metric algebra, a logical foundation for embedding-independence. It models each dataset as a metric relation, i.e., identifiers equipped with a semantic distance, and provides operators for similarity search, and, critically, a semantics-preserving Union operator for cross-embedding search. We implement metric algebra in MetricDB, which treats existing vector databases as physical vector realizations that implement metric relations, and executes Union via witness maps learned from overlap anchors. Building on this logical-physical bridge, MetricDB plans hard vs. soft execution for queries over Union under imperfect witnesses and avoids rebuilds via transformation-equivariant index compilation and merging. Experiments on benchmark workloads across heterogeneous embedding models (e.g., Mistral, GTE-Qwen2, NV-Embed-V2, OpenAI-Ada-002) demonstrate robust cross-embedding retrieval and efficient interoperability.
- 24 September 2026
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Management of Data
- Vol. 4 (4) , 1-26
Open AccessMetric Algebra: Embedding-Independence in Vector Databases
Vector databases are increasingly central to retrieval applications, yet deployments remain embedding-siloed: different databases store vectors produced by different embedding models and ANN pipelines, making cross-database search ill-defined or forcing expensive re-embedding and index rebuilds. We propose metric algebra, a logical foundation for embedding-independence. It models each dataset as a metric relation, i.e., identifiers equipped with a semantic distance, and provides operators for similarity search, and, critically, a semantics-preserving Union operator for cross-embedding search. We implement metric algebra in MetricDB, which treats existing vector databases as physical vector realizations that implement metric relations, and executes Union via witness maps learned from overlap anchors. Building on this logical-physical bridge, MetricDB plans hard vs. soft execution for queries over Union under imperfect witnesses and avoids rebuilds via transformation-equivariant index compilation and merging. Experiments on benchmark workloads across heterogeneous embedding models (e.g., Mistral, GTE-Qwen2, NV-Embed-V2, OpenAI-Ada-002) demonstrate robust cross-embedding retrieval and efficient interoperability.
- Open Access
- 20 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 1554-1565
Mitigating Privacy Risks in Graph Condensation from a Hyperbolic Geometry Perspective
- 20 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 1554-1565
Open AccessMitigating Privacy Risks in Graph Condensation from a Hyperbolic Geometry Perspective
- Open Access
- 4 December 2025
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Management of Data
- Vol. 3 (6) , 1-28
Integrating Vector Databases across Embedding Models
Vector databases have been widely used to implement similarity search over unstructured objects, e.g., documents and images. Each vector database is produced by an embedding model that encodes the objects in a way such that more similar objects are embedded to closer vectors, allowing us to use top-k vector search as an implementation of top-k object similarity search. It is common practice that different vector databases use distinct embedding models and the same object may be encoded by different embedding vectors across databases. As a result, one cannot share and integrate vector databases to expand similarity search across datasets, a property we take for granted for relational databases. In this work, we attempt to break the barrier between different vector databases, by developing an approach to integrating vector databases generated by different embedding models, with neither any access to the encoded data objects nor knowledge of the embedding models. Our approach is rooted in the local isometry hypothesis, a finding made via extensive experiments on real-life embedding vectors, and is backed up by theoretical analysis that bounds the quality of integrated vector database. Experimental results show that we can integrate vector databases produced by various popular embedding models, e.g., NV-embed-V2, OpenAI Ada, GloVe, Mistral and FastText, while offering high recall of top-k similarity search over the integrated datasets.
- 4 December 2025
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Management of Data
- Vol. 3 (6) , 1-28
Open AccessIntegrating Vector Databases across Embedding Models
Vector databases have been widely used to implement similarity search over unstructured objects, e.g., documents and images. Each vector database is produced by an embedding model that encodes the objects in a way such that more similar objects are embedded to closer vectors, allowing us to use top-k vector search as an implementation of top-k object similarity search. It is common practice that different vector databases use distinct embedding models and the same object may be encoded by different embedding vectors across databases. As a result, one cannot share and integrate vector databases to expand similarity search across datasets, a property we take for granted for relational databases. In this work, we attempt to break the barrier between different vector databases, by developing an approach to integrating vector databases generated by different embedding models, with neither any access to the encoded data objects nor knowledge of the embedding models. Our approach is rooted in the local isometry hypothesis, a finding made via extensive experiments on real-life embedding vectors, and is backed up by theoretical analysis that bounds the quality of integrated vector database. Experimental results show that we can integrate vector databases produced by various popular embedding models, e.g., NV-embed-V2, OpenAI Ada, GloVe, Mistral and FastText, while offering high recall of top-k similarity search over the integrated datasets.
- Open Access
- 24 August 2024
- Published by Association for Computing Machinery (ACM)
- p. 4467-4478
Pre-Training and Prompting for Few-Shot Node Classification on Text-Attributed Graphs
The text-attributed graph (TAG) is one kind of important real-world graph-structured data with each node associated with raw texts. For TAGs, traditional few-shot node classification methods directly conduct training on the pre-processed node features and do not consider the raw texts. The performance is highly dependent on the choice of the feature pre-processing method. In this paper, we propose P2TAG, a framework designed for few-shot node classification on TAGs with graph pre-training and prompting. P2TAG first pre-trains the language model (LM) and graph neural network (GNN) on TAGs with self-supervised loss. To fully utilize the ability of language models, we adapt the masked language modeling objective for our framework. The pre-trained model is then used for the few-shot node classification with a mixed prompt method, which simultaneously considers both text and graph information. We conduct experiments on six real-world TAGs, including paper citation networks and product co-purchasing networks. Experimental results demonstrate that our proposed framework outperforms existing graph few-shot learning methods on these datasets with +18.98% ~ +32.14% improvements.
- 24 August 2024
- Published by Association for Computing Machinery (ACM)
- p. 4467-4478
Open AccessPre-Training and Prompting for Few-Shot Node Classification on Text-Attributed Graphs
The text-attributed graph (TAG) is one kind of important real-world graph-structured data with each node associated with raw texts. For TAGs, traditional few-shot node classification methods directly conduct training on the pre-processed node features and do not consider the raw texts. The performance is highly dependent on the choice of the feature pre-processing method. In this paper, we propose P2TAG, a framework designed for few-shot node classification on TAGs with graph pre-training and prompting. P2TAG first pre-trains the language model (LM) and graph neural network (GNN) on TAGs with self-supervised loss. To fully utilize the ability of language models, we adapt the masked language modeling objective for our framework. The pre-trained model is then used for the few-shot node classification with a mixed prompt method, which simultaneously considers both text and graph information. We conduct experiments on six real-world TAGs, including paper citation networks and product co-purchasing networks. Experimental results demonstrate that our proposed framework outperforms existing graph few-shot learning methods on these datasets with +18.98% ~ +32.14% improvements.