Zhihuan Huang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Zhihuan Huang has a total of 10 co-authors. Here are the top co-authors by number of shared publications:
- Yongkang Guo
School of Computer Science, Peking University, Beijing, China
2 shared publications - Yuqing Kong
CFCS, School of Computer Science, Peking University, Beijing, China
2 shared publications - Anant Shah
Northwestern University, Evanston, USA
1 shared publication - Fang-Yi Yu
George Mason University
1 shared publication - Jason D. Hartline
Computer Science, Northwestern University, Evanston, IL, USA
1 shared publication - Michail Iosifovich Kogan
Bowers Neurosurgical Frailty and Outcomes Data Science Lab, Flint, Michigan, USA
1 shared publication
Latest Publications
- Open Access
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 99-110
PMIScore: An Unsupervised Approach to Quantify Dialogue Engagement
High dialogue engagement is a crucial indicator of an effective conversation. A reliable measure of engagement could help benchmark large language models, enhance the effectiveness of human-computer interactions, or improve personal communication skills. However, quantifying engagement is challenging, since it is subjective and lacks a ''gold standard''. This paper proposes PMIScore, an efficient unsupervised approach to quantify dialogue engagement. It uses pointwise mutual information (PMI), which is the probability of generating a response conditioning on the conversation history. Thus, PMIScore offers a clear interpretation of engagement. As directly computing PMI is intractable due to the complexity of dialogues, PMIScore learned it through a dual form of divergence. The algorithm includes generating positive and negative dialogue pairs, extracting embeddings by large language models (LLMs), and training a small neural network using a mutual information loss function. We validated PMIScore on both synthetic and real-world datasets. Our results demonstrate the effectiveness of PMIScore in PMI estimation and the reasonableness of the PMI metric itself.
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 99-110
Open AccessPMIScore: An Unsupervised Approach to Quantify Dialogue Engagement
High dialogue engagement is a crucial indicator of an effective conversation. A reliable measure of engagement could help benchmark large language models, enhance the effectiveness of human-computer interactions, or improve personal communication skills. However, quantifying engagement is challenging, since it is subjective and lacks a ''gold standard''. This paper proposes PMIScore, an efficient unsupervised approach to quantify dialogue engagement. It uses pointwise mutual information (PMI), which is the probability of generating a response conditioning on the conversation history. Thus, PMIScore offers a clear interpretation of engagement. As directly computing PMI is intractable due to the complexity of dialogues, PMIScore learned it through a dual form of divergence. The algorithm includes generating positive and negative dialogue pairs, extracting embeddings by large language models (LLMs), and training a small neural network using a mutual information loss function. We validated PMIScore on both synthetic and real-world datasets. Our results demonstrate the effectiveness of PMIScore in PMI estimation and the reasonableness of the PMI metric itself.
- Open Access
- 2 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 1110-1129
Algorithmic Robust Forecast Aggregation
- 2 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 1110-1129
Open AccessAlgorithmic Robust Forecast Aggregation
- Open Access
- 15 March 2017
- Published by Wiley in American Journal of Botany
- Vol. 104 (3) , 468-476
Nectar properties and the role of sunbirds as pollinators of the golden‐flowered tea (Camellia petelotii)
PREMISE OF THE STUDY: Properties of floral nectar have been used to predict if a plant species is pollinated by birds. To see whether winter‐flowering plants evolve nectar properties corresponding to bird pollinators, nectar properties of several Camellia species (including the golden‐flowered tea), as well as the role of floral visitors as effective pollinators, were examined.METHODS: Potential pollinators of Camellia petelotii were identified at different times of day and under various weather conditions. A bird exclusion experiment was used to compare the pollination effectiveness of birds and insects. Nectar sugar components (fructose, glucose, and sucrose) from C. petelotii growing wild and another seven Camellia species and 22 additional cultivars (all in cultivation) were examined by high‐performance liquid chromatography (HPLC).KEY RESULTS: The sunbird Aethopyga siparaja and honeybees were the most frequent floral visitors to C. petelotii. Honeybee visits were significantly reduced in cloudy/rainy weather. The fruit and seed set of flowers with birds excluded were reduced by 64%, indicating that bird pollination is significant. For the wild populations of C. petelotii, a bagged flower could secrete 157 μL nectar; this nectar has a low sugar concentration (19%) and is sucrose‐dominant (87%). The eight Camellia species and 22 cultivars had an average sugar concentration of around 30% and a sucrose concentration of 80%, demonstrating sucrose‐dominant nectar in Camellia species.CONCLUSIONS: The nectar sugar composition of Camellia species was characterized by sucrose dominance. In addition, the large reduction in seed set when birds are excluded in the golden‐flowered tea also supports the suggestion that these winter‐flowering plants may have evolved with birds as significant pollinators.
- 15 March 2017
- Published by Wiley in American Journal of Botany
- Vol. 104 (3) , 468-476
Open AccessNectar properties and the role of sunbirds as pollinators of the golden‐flowered tea (Camellia petelotii)
PREMISE OF THE STUDY: Properties of floral nectar have been used to predict if a plant species is pollinated by birds. To see whether winter‐flowering plants evolve nectar properties corresponding to bird pollinators, nectar properties of several Camellia species (including the golden‐flowered tea), as well as the role of floral visitors as effective pollinators, were examined.METHODS: Potential pollinators of Camellia petelotii were identified at different times of day and under various weather conditions. A bird exclusion experiment was used to compare the pollination effectiveness of birds and insects. Nectar sugar components (fructose, glucose, and sucrose) from C. petelotii growing wild and another seven Camellia species and 22 additional cultivars (all in cultivation) were examined by high‐performance liquid chromatography (HPLC).KEY RESULTS: The sunbird Aethopyga siparaja and honeybees were the most frequent floral visitors to C. petelotii. Honeybee visits were significantly reduced in cloudy/rainy weather. The fruit and seed set of flowers with birds excluded were reduced by 64%, indicating that bird pollination is significant. For the wild populations of C. petelotii, a bagged flower could secrete 157 μL nectar; this nectar has a low sugar concentration (19%) and is sucrose‐dominant (87%). The eight Camellia species and 22 cultivars had an average sugar concentration of around 30% and a sucrose concentration of 80%, demonstrating sucrose‐dominant nectar in Camellia species.CONCLUSIONS: The nectar sugar composition of Camellia species was characterized by sucrose dominance. In addition, the large reduction in seed set when birds are excluded in the golden‐flowered tea also supports the suggestion that these winter‐flowering plants may have evolved with birds as significant pollinators.
- Open Access
- 1 November 2008
- Published by American Public Health Association in American Journal of Public Health
- Vol. 98 (11) , 1996-2003
State-Level Health Care Access and Use Among Children in US Immigrant Families
Objectives. We examined the association between children's state of residence and their access to health care among specific types of immigrant families: foreign-born children, US-born children with 1 foreign-born parent, US-born children with both foreign-born parents, and nonimmigrant families. Methods. We analyzed data from 12 400 children from the 2003 National Survey of Children's Health in the 6 states with the highest proportion of immigrants (California, Florida, Illinois, New York, New Jersey, and Texas). Results. Multivariable analyses indicated that among foreign-born children, those living in California, Illinois, and Texas were more likely to lack access to health care compared with those living in New York. Among foreign-born children with 1 or 2 US-born parents, Texas children were most likely to lack health insurance. Within nonimmigrant families, children from California, Florida, and Texas had significantly more access and use problems. Conclusions. Our findings document differential health care access and use among states for specific immigrant family types.
- 1 November 2008
- Published by American Public Health Association in American Journal of Public Health
- Vol. 98 (11) , 1996-2003
Open AccessState-Level Health Care Access and Use Among Children in US Immigrant Families
Objectives. We examined the association between children's state of residence and their access to health care among specific types of immigrant families: foreign-born children, US-born children with 1 foreign-born parent, US-born children with both foreign-born parents, and nonimmigrant families. Methods. We analyzed data from 12 400 children from the 2003 National Survey of Children's Health in the 6 states with the highest proportion of immigrants (California, Florida, Illinois, New York, New Jersey, and Texas). Results. Multivariable analyses indicated that among foreign-born children, those living in California, Illinois, and Texas were more likely to lack access to health care compared with those living in New York. Among foreign-born children with 1 or 2 US-born parents, Texas children were most likely to lack health insurance. Within nonimmigrant families, children from California, Florida, and Texas had significantly more access and use problems. Conclusions. Our findings document differential health care access and use among states for specific immigrant family types.