Tongran Liu
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
Tongran Liu has a total of 42 co-authors. Here are the top co-authors by number of shared publications:
Jiannong Shi
CAS Key Laboratory of Behavioral Science, Institute of Psychology, Chinese Academy of Sciences, Beijing, China; Department of Psychology, University of Chinese Academy of Sciences, Beijing, China; Department of Learning and Philosophy, Aalborg University, Denmark
10 shared publications- Jiannong Shi
Chinese Academy of Sciences
7 shared publications Tong Xiao
School of Computer Science and Engineering, Northeastern University, Shenyang, China
6 shared publicationsDanfeng Li
CAS Key Laboratory of Behavioral Science, Institute of Psychology, Chinese Academy of Sciences, Beijing 100101, China; Department of Psychology, University of Chinese Academy of Sciences, Beijing 100049, China; School of Sociology and Psychology, Central University of Finance and Economics, Beijing, China
4 shared publicationsDi Wang
CAS Key Laboratory of Behavioral Science, Institute of Psychology, Chinese Academy of Sciences, Beijing, China; Department of Psychology, University of Chinese Academy of Sciences, Beijing, China
3 shared publicationsFangfang Shangguan
Beijing Key Laboratory of Learning and Cognition, School of Psychology, Capital Normal University, Beijing, China
3 shared publications
Latest Publications
- Open Access
- 19 February 2026
- Published by Elsevier in NeuroImage
- Vol. 329, 121818
The effects of working memory and interference control on reinforcement learning: Evidence from computational modeling and structural MRI
- 19 February 2026
- Published by Elsevier in NeuroImage
- Vol. 329, 121818
Open AccessThe effects of working memory and interference control on reinforcement learning: Evidence from computational modeling and structural MRI
- Open Access
- 16 October 2025
- Published by Springer Nature in npj Science of Learning
- Vol. 10 (1) , 1-13
The distinct functions of working memory and intelligence in model-based and model-free reinforcement learning
Human and animal behaviors are influenced by goal-directed planning or automatic habitual choices. Reinforcement learning (RL) models propose two distinct learning strategies: a model-based strategy, which is more flexible but computationally demanding, and a model-free strategy is less flexible yet computationally efficient. In the current RL tasks, we investigated how individuals adjusted these strategies under varying working memory (WM) loads and further explored how learning strategies and mental abilities (WM capacity and intelligence) affected learning performance. The results indicated that participants were more inclined to employ the model-based strategy under low WM load, while shifting towards the model-free strategy under high WM load. Linear regression models suggested that the utilization of model-based strategy and intelligence positively predicted learning performance. Furthermore, the model-based learning strategy could mediate the influence of WM load on learning performance. These findings underscore the critical role of WM capacity in strategic selection during RL process.
- 16 October 2025
- Published by Springer Nature in npj Science of Learning
- Vol. 10 (1) , 1-13
Open AccessThe distinct functions of working memory and intelligence in model-based and model-free reinforcement learning
Human and animal behaviors are influenced by goal-directed planning or automatic habitual choices. Reinforcement learning (RL) models propose two distinct learning strategies: a model-based strategy, which is more flexible but computationally demanding, and a model-free strategy is less flexible yet computationally efficient. In the current RL tasks, we investigated how individuals adjusted these strategies under varying working memory (WM) loads and further explored how learning strategies and mental abilities (WM capacity and intelligence) affected learning performance. The results indicated that participants were more inclined to employ the model-based strategy under low WM load, while shifting towards the model-free strategy under high WM load. Linear regression models suggested that the utilization of model-based strategy and intelligence positively predicted learning performance. Furthermore, the model-based learning strategy could mediate the influence of WM load on learning performance. These findings underscore the critical role of WM capacity in strategic selection during RL process.
- Open Access
- 24 May 2022
- Published by Elsevier in Developmental Cognitive Neuroscience
- Vol. 55, 101118
The influence of reward anticipation on conflict control in children and adolescents: Evidences from hierarchical drift-diffusion model and event-related potentials
- 24 May 2022
- Published by Elsevier in Developmental Cognitive Neuroscience
- Vol. 55, 101118
Open AccessThe influence of reward anticipation on conflict control in children and adolescents: Evidences from hierarchical drift-diffusion model and event-related potentials
- Open Access
- 16 April 2021
- Published by Springer Nature in Scientific Reports
- Vol. 11 (1) , 1-18
The neurodevelopment of delay discounting for monetary rewards in pre-adolescent children
Children are found to exhibit high degrees of delay discounting compared with adults in many delay discounting studies, which might be due to the asynchronous development of “bottom-up” and “top-down” neural systems. However, the temporal dynamics associated with the two systems in the development of delay discounting processes are not well known. In this study, we chose two age groups of participants and adopted event-related potential (ERP) techniques to investigate the neural dynamic differences between children and adults during delay discounting processes. Behavioral findings showed that children discounted more than adults and chose more immediate choices. Electrophysiological findings revealed that children exhibited longer neural processing (longer P2 latency) than adults during the early detection and identification phase. Children showed less cognitive control (smaller N2 amplitude) than adults over the middle frontal areas, and they devoted more neural effort (larger P3 amplitudes) to making final choices than adults. The factors of reward amount and time delay could influence the development of delay discounting in children.
- 16 April 2021
- Published by Springer Nature in Scientific Reports
- Vol. 11 (1) , 1-18
Open AccessThe neurodevelopment of delay discounting for monetary rewards in pre-adolescent children
Children are found to exhibit high degrees of delay discounting compared with adults in many delay discounting studies, which might be due to the asynchronous development of “bottom-up” and “top-down” neural systems. However, the temporal dynamics associated with the two systems in the development of delay discounting processes are not well known. In this study, we chose two age groups of participants and adopted event-related potential (ERP) techniques to investigate the neural dynamic differences between children and adults during delay discounting processes. Behavioral findings showed that children discounted more than adults and chose more immediate choices. Electrophysiological findings revealed that children exhibited longer neural processing (longer P2 latency) than adults during the early detection and identification phase. Children showed less cognitive control (smaller N2 amplitude) than adults over the middle frontal areas, and they devoted more neural effort (larger P3 amplitudes) to making final choices than adults. The factors of reward amount and time delay could influence the development of delay discounting in children.
- 1 September 2020
- Published by Elsevier in International Journal of Psychophysiology
- Vol. 155, 8
Corrigendum to “Fluid intelligence and neural mechanisms of emotional conflict adaptation” [Int. J. Psychophysiol. 152 (2020) 1–14]
- 1 September 2020
- Published by Elsevier in International Journal of Psychophysiology
- Vol. 155, 8
Corrigendum to “Fluid intelligence and neural mechanisms of emotional conflict adaptation” [Int. J. Psychophysiol. 152 (2020) 1–14]
- Open Access
The neurodevelopment of delay discounting for monetary rewards in pre-adolescent children
Children are found to exhibit high degrees of delay discounting compared with adults in many delay discounting studies. However, the temporal dynamics of the differences behind the behaviors are not well known. In this study, we chose two age groups of participants and adopted event-related potential (ERP) techniques to investigate the neural dynamic differences between children and adults during delay discounting processes. Behavioral findings showed that children discounted more than adults and chose more immediate choices. Electrophysiological findings revealed that children exhibited longer neural processing (longer P2 latency) than adults during the early detection and identification phase. Children showed less cognitive control (smaller N2 amplitude) than adults over the middle frontal areas, and they devoted more neural effort (larger P3 amplitudes) to making final choices than adults. The factors of reward amount and time delay could influence the development of delay discounting in male children.
Open AccessThe neurodevelopment of delay discounting for monetary rewards in pre-adolescent children
Children are found to exhibit high degrees of delay discounting compared with adults in many delay discounting studies. However, the temporal dynamics of the differences behind the behaviors are not well known. In this study, we chose two age groups of participants and adopted event-related potential (ERP) techniques to investigate the neural dynamic differences between children and adults during delay discounting processes. Behavioral findings showed that children discounted more than adults and chose more immediate choices. Electrophysiological findings revealed that children exhibited longer neural processing (longer P2 latency) than adults during the early detection and identification phase. Children showed less cognitive control (smaller N2 amplitude) than adults over the middle frontal areas, and they devoted more neural effort (larger P3 amplitudes) to making final choices than adults. The factors of reward amount and time delay could influence the development of delay discounting in male children.
- Open Access
Learning Architectures from an Extended Search Space for Language Modeling
Neural architecture search (NAS) has advanced significantly in recent years but most NAS systems restrict search to learning architectures of a recurrent or convolutional cell. In this paper, we extend the search space of NAS. In particular, we present a general approach to learn both intra-cell and inter-cell architectures (call it ESS). For a better search result, we design a joint learning method to perform intra-cell and inter-cell NAS simultaneously. We implement our model in a differentiable architecture search system. For recurrent neural language modeling, it outperforms a strong baseline significantly on the PTB and WikiText data, with a new state-of-the-art on PTB. Moreover, the learned architectures show good transferability to other systems. E.g., they improve state-of-the-art systems on the CoNLL and WNUT named entity recognition (NER) tasks and CoNLL chunking task, indicating a promising line of research on large-scale pre-learned architectures.
Open AccessLearning Architectures from an Extended Search Space for Language Modeling
Neural architecture search (NAS) has advanced significantly in recent years but most NAS systems restrict search to learning architectures of a recurrent or convolutional cell. In this paper, we extend the search space of NAS. In particular, we present a general approach to learn both intra-cell and inter-cell architectures (call it ESS). For a better search result, we design a joint learning method to perform intra-cell and inter-cell NAS simultaneously. We implement our model in a differentiable architecture search system. For recurrent neural language modeling, it outperforms a strong baseline significantly on the PTB and WikiText data, with a new state-of-the-art on PTB. Moreover, the learned architectures show good transferability to other systems. E.g., they improve state-of-the-art systems on the CoNLL and WNUT named entity recognition (NER) tasks and CoNLL chunking task, indicating a promising line of research on large-scale pre-learned architectures.
- 1 June 2020
- Published by Elsevier in International Journal of Psychophysiology
- Vol. 152, 1-14
Fluid intelligence and neural mechanisms of emotional conflict adaptation
- 1 June 2020
- Published by Elsevier in International Journal of Psychophysiology
- Vol. 152, 1-14
Fluid intelligence and neural mechanisms of emotional conflict adaptation
- Open Access
Does Multi-Encoder Help? A Case Study on Context-Aware Neural Machine Translation
In encoder-decoder neural models, multiple encoders are in general used to represent the contextual information in addition to the individual sentence. In this paper, we investigate multi-encoder approaches in documentlevel neural machine translation (NMT). Surprisingly, we find that the context encoder does not only encode the surrounding sentences but also behaves as a noise generator. This makes us rethink the real benefits of multi-encoder in context-aware translation - some of the improvements come from robust training. We compare several methods that introduce noise and/or well-tuned dropout setup into the training of these encoders. Experimental results show that noisy training plays an important role in multi-encoder-based NMT, especially when the training data is small. Also, we establish a new state-of-the-art on IWSLT Fr-En task by careful use of noise generation and dropout methods.
Open AccessDoes Multi-Encoder Help? A Case Study on Context-Aware Neural Machine Translation
In encoder-decoder neural models, multiple encoders are in general used to represent the contextual information in addition to the individual sentence. In this paper, we investigate multi-encoder approaches in documentlevel neural machine translation (NMT). Surprisingly, we find that the context encoder does not only encode the surrounding sentences but also behaves as a noise generator. This makes us rethink the real benefits of multi-encoder in context-aware translation - some of the improvements come from robust training. We compare several methods that introduce noise and/or well-tuned dropout setup into the training of these encoders. Experimental results show that noisy training plays an important role in multi-encoder-based NMT, especially when the training data is small. Also, we establish a new state-of-the-art on IWSLT Fr-En task by careful use of noise generation and dropout methods.
- 1 May 2020
- Published by Elsevier in Neuropsychologia
- Vol. 142, 107469
Food is discounted more steeply than money: Evidence from N2 and P3 responses in delay discounting tasks
- 1 May 2020
- Published by Elsevier in Neuropsychologia
- Vol. 142, 107469
Food is discounted more steeply than money: Evidence from N2 and P3 responses in delay discounting tasks