Xiao Yang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Xiao Yang has a total of 53 co-authors. Here are the top co-authors by number of shared publications:
- Kai Sun
Meta, Menlo Park, CA, USA
3 shared publications - Qi Li
State Key Laboratory of Power Systems, Department of Electrical Engineering, Tsinghua University, Beijing, China
2 shared publications - Sanat Sharma
Meta, Menlo Park, CA, USA
2 shared publications - Wen-Tau Yih
Meta, Menlo Park, CA, USA
2 shared publications - Xiaolei Zhao
School of Electrical Engineering, Shanghai Jiao Tong University (SJTU), Shanghai, China
2 shared publications - Xin Luna Dong
Meta Reality Labs, Redmond, Washington, USA
2 shared publications
Latest Publications
- Open Access
- 8 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 13308-13313
Retrieval-Augmented Generation (RAG)— From Modular to Agentic Systems
- 8 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 13308-13313
Open AccessRetrieval-Augmented Generation (RAG)— From Modular to Agentic Systems
- Open Access
- 8 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 9906-9915
CRAG-MM: Multi-modal Multi-turn Comprehensive RAG Benchmark
- 8 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 9906-9915
Open AccessCRAG-MM: Multi-modal Multi-turn Comprehensive RAG Benchmark
- 17 March 2025
- Published by IOS Press
Knowledge Internalized in LLMs
LLMs, trained on a diverse range of resources, internalize a substantial amount of knowledge. Understanding this internalized knowledge is crucial for assessing the accuracy and reliability of their outputs, guiding improvements in model design, and addressing potential biases. In this chapter, we explore two primary forms of knowledge within LLMs: taxonomy and factual knowledge. We delve into the nature of internalized knowledge, examining how these models store and utilize information. Additionally, we introduce CRAG, a comprehensive benchmark for evaluating both the internalized knowledge of LLMs and their performance in retrieved-augmented generation tasks.
- 17 March 2025
- Published by IOS Press
Knowledge Internalized in LLMs
LLMs, trained on a diverse range of resources, internalize a substantial amount of knowledge. Understanding this internalized knowledge is crucial for assessing the accuracy and reliability of their outputs, guiding improvements in model design, and addressing potential biases. In this chapter, we explore two primary forms of knowledge within LLMs: taxonomy and factual knowledge. We delve into the nature of internalized knowledge, examining how these models store and utilize information. Additionally, we introduce CRAG, a comprehensive benchmark for evaluating both the internalized knowledge of LLMs and their performance in retrieved-augmented generation tasks.
- 31 October 2019
- Published by Springer Nature
- p. 52-60
Design and Electric Field Calculation of a Wall Bushing Made from Nonlinear Materials
- 31 October 2019
- Published by Springer Nature
- p. 52-60
Design and Electric Field Calculation of a Wall Bushing Made from Nonlinear Materials
- 19 April 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Dielectrics and Electrical Insulation
- Vol. 25 (2) , 623-630
How nonlinear V-I characteristics of single ZnO microvaristor influences the performance of its silicone rubber composite
Microvaristor ceramics blended with insulation polymers can function as nonlinear dielectrics, which possess self-adaptive electrical parameters varying with the applied field. Broadening the switching (electric) field of the nonlinear dielectrics for their applications in various apparatus and voltage class is an important issue. This paper prepared ZnO microvaristor ceramics with different grain size through controlling the sintering temperature, as well as the microvaristor ceramic/polymer composites (MCPC) with different filler concentration, filler diameter, and filler grain size. The nonlinear V-I characteristics of both the single ZnO microvaristor filler and the MCPC were measured. The switching field of the MCPC is found to be adjusted in a wider range by controlling the filler grain size. The relationship between the nonlinear conducting behavior of single filler and the composite are analyzed. Suggestions were proposed for the preparation of ZnO microvaristor fillers to meet specific field grading requirement for MCPCs.
- 19 April 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Dielectrics and Electrical Insulation
- Vol. 25 (2) , 623-630
How nonlinear V-I characteristics of single ZnO microvaristor influences the performance of its silicone rubber composite
Microvaristor ceramics blended with insulation polymers can function as nonlinear dielectrics, which possess self-adaptive electrical parameters varying with the applied field. Broadening the switching (electric) field of the nonlinear dielectrics for their applications in various apparatus and voltage class is an important issue. This paper prepared ZnO microvaristor ceramics with different grain size through controlling the sintering temperature, as well as the microvaristor ceramic/polymer composites (MCPC) with different filler concentration, filler diameter, and filler grain size. The nonlinear V-I characteristics of both the single ZnO microvaristor filler and the MCPC were measured. The switching field of the MCPC is found to be adjusted in a wider range by controlling the filler grain size. The relationship between the nonlinear conducting behavior of single filler and the composite are analyzed. Suggestions were proposed for the preparation of ZnO microvaristor fillers to meet specific field grading requirement for MCPCs.
- 1 February 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Information Theory
- Vol. 63 (00189448) , 1
Large Alphabet Compression and Predictive Distributions through Poissonization and Tilting
This paper introduces a convenient strategy for coding and predicting sequences of independent, identically distributed random variables generated from a large alphabet of size m. In particular, the size of the sample is allowed to be variable. The employment of a Poisson model and tilting method simplifies the implementation and analysis through independence. The resulting strategy is optimal within the class of distributions satisfying a moment condition, and it is close to optimal for the class of all i.i.d distributions on strings of a given length. The method also can be used to code and predict strings with a condition on the tail of the ordered counts, and it can be applied to distributions in an envelope class. Moreover, we show that our model permits exact computation of the minimax optimal code, for all alphabet sizes, when conditioning on the size of the sample.
- 1 February 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Information Theory
- Vol. 63 (00189448) , 1
Large Alphabet Compression and Predictive Distributions through Poissonization and Tilting
This paper introduces a convenient strategy for coding and predicting sequences of independent, identically distributed random variables generated from a large alphabet of size m. In particular, the size of the sample is allowed to be variable. The employment of a Poisson model and tilting method simplifies the implementation and analysis through independence. The resulting strategy is optimal within the class of distributions satisfying a moment condition, and it is close to optimal for the class of all i.i.d distributions on strings of a given length. The method also can be used to code and predict strings with a condition on the tail of the ordered counts, and it can be applied to distributions in an envelope class. Moreover, we show that our model permits exact computation of the minimax optimal code, for all alphabet sizes, when conditioning on the size of the sample.