Author name cluster

Junjie Liu

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

6 papers

2 author rows

EAAI Journal 2026 Journal Article

Intelligent attitude control of fighter aircraft at high angle of attack based on predefined-time observation and deep reinforcement learning

Junjie Liu
Yetong Lin
Yuehui Ji
Yu Song
Qiang Gao

To address the challenges of low parameter tuning efficiency and insufficient disturbance rejection performance in traditional active disturbance rejection control (ADRC) under the highly nonlinear and strongly coupled dynamics of fighter aircraft at high angle of attack, this paper proposes an intelligent attitude control method that integrates deep reinforcement learning with predefined-time state observation. First, a predefined-time super-twisting extended state observer (PTSTESO) is designed within the ADRC framework to improve real-time estimation accuracy of total disturbances through a convergence mechanism that is independent of initial conditions. Then, a coordinated optimization framework based on a dual-enhanced twin-delayed deep deterministic policy gradient (TD3) algorithm is developed. In this framework, the first agent employs a network structure combining gated recurrent unit (GRU) and a self attention (SA) mechanism to capture temporal dependencies and highlight critical features in the state sequence, thereby enabling adaptive and precise tuning of ADRC parameters. The second agent integrates prioritized experience replay (PER) and a novel action exploration strategy to improve sampling efficiency and learning performance, enabling effective control law optimization in the angle-of-attack channel. Simulation results show that the proposed method effectively reduces manual controller tuning effort and improves tracking accuracy and system robustness. Under a representative high-angle-of-attack maneuvering condition, the proposed method reduces the integral absolute error (IAE) and the integral time-weighted absolute error (ITAE) of angle-of-attack tracking by 96. 66% and 96. 71% compared with the conventional active disturbance rejection control, and by 94. 80% and 95. 06% compared with the modified active disturbance rejection control, respectively.

Details DOI

NeurIPS Conference 2025 Conference Paper

GeoCAD: Local Geometry-Controllable CAD Generation with Large Language Models

Zhanwei Zhang
Kaiyuan Liu
Junjie Liu
Wenxiao Wang
Binbin Lin
Liang Xie
Chen Shen
Deng Cai

Local geometry-controllable computer-aided design (CAD) generation aims to modify local parts of CAD models automatically, enhancing design efficiency. It also ensures that the shapes of newly generated local parts follow user-specific geometric instructions (e. g. , an isosceles right triangle or a rectangle with one corner cut off). However, existing methods encounter challenges in achieving this goal. Specifically, they either lack the ability to follow textual instructions or are unable to focus on the local parts. To address this limitation, we introduce GeoCAD, a user-friendly and local geometry-controllable CAD generation method. Specifically, we first propose a complementary captioning strategy to generate geometric instructions for local parts. This strategy involves vertex-based and VLLM-based captioning for systematically annotating simple and complex parts, respectively. In this way, we caption $\sim$221k different local parts in total. In the training stage, given a CAD model, we randomly mask a local part. Then, using its geometric instruction and the remaining parts as input, we prompt large language models (LLMs) to predict the masked part. During inference, users can specify any local part for modification while adhering to a variety of predefined geometric instructions. Extensive experiments demonstrate the effectiveness of GeoCAD in generation quality, validity and text-to-CAD consistency.

PDF Details

EAAI Journal 2024 Journal Article

Multiple prior representation learning for self-supervised monocular depth estimation via hybrid transformer

Guodong Sun
Junjie Liu
Mingxuan Liu
Moyun Liu
Yang Zhang

Self-supervised monocular depth estimation aims to infer depth information without relying on labeled data. However, the lack of labeled information poses a significant challenge to the model’s representation, limiting its ability to capture the intricate details of the scene accurately. Prior information can potentially mitigate this issue, enhancing the model’s understanding of scene structure and texture. Nevertheless, solely relying on a single type of prior information often falls short when dealing with complex scenes, necessitating improvements in generalization performance. To address these challenges, we introduce a novel self-supervised monocular depth estimation model that leverages multiple priors to bolster representation capabilities across spatial, context, and semantic dimensions. Specifically, we employ a hybrid transformer and a lightweight pose network to obtain long-range spatial priors in the spatial dimension. Then, the context prior attention is designed to improve generalization, particularly in complex structures or untextured areas. In addition, semantic priors are introduced by leveraging semantic boundary loss, and semantic prior attention is supplemented, further refining the semantic features extracted by the decoder. Experiments on three diverse datasets demonstrate the effectiveness of the proposed model. It integrates multiple priors to comprehensively enhance the representation ability, improving the accuracy and reliability of depth estimation. Codes are available at: https: //github. com/MVME-HBUT/MPRLNet.

Details DOI

JBHI Journal 2022 Journal Article

Investigating of Deaf Emotion Cognition Pattern By EEG and Facial Expression Combination

Yi Yang
Qiang Gao
Yu Song
Xiaolin Song
Zemin Mao
Junjie Liu

With the development of sensor technology and learning algorithms, multimodal emotion recognition has attracted widespread attention. Many existing studies on emotion recognition mainly focused on normal people. Besides, due to hearing loss, deaf people cannot express emotions by words, which may have a greater need for emotion recognition. In this paper, the deep belief network (DBN) was utilized to classify three category emotions through the electroencephalograph (EEG) and facial expressions. Signals from 15 deaf subjects were recorded when they watched the emotional movie clips. Our system uses a 1-s window without overlap to segment the EEG signals in five frequency bands, then the differential entropy (DE) feature is extracted. The DE feature of EEG and facial expression images plays as multimodal input for subject-dependent emotion recognition. To avoid feature redundancy, the top 12 major EEG electrode channels (FP2, FP1, FT7, FPZ, F7, T8, F8, CB2, CB1, FT8, T7, TP8) in the gamma band and 30 facial expression features (the areas around the eyes and eyebrow) which are selected by the largest weight values. The results show that the classification accuracy is 99. 92% by feature selection in deaf emotion reignition. Moreover, investigations on brain activities reveal deaf brain activity changes mainly in the beta and gamma bands, and the brain regions that are affected by emotions are mainly distributed in the prefrontal and outer temporal lobes.

Details DOI

ICLR Conference 2020 Conference Paper

Dynamic Sparse Training: Find Efficient Sparse Network From Scratch With Trainable Masked Layers

Junjie Liu
Zhe Xu 0008
Runbin Shi
Ray C. C. Cheung
Hayden Kwok-Hay So

We present a novel network pruning algorithm called Dynamic Sparse Training that can jointly ﬁnd the optimal network parameters and sparse network structure in a uniﬁed optimization process with trainable pruning thresholds. These thresholds can have ﬁne-grained layer-wise adjustments dynamically via backpropagation. We demonstrate that our dynamic sparse training algorithm can easily train very sparse neural network models with little performance loss using the same training epochs as dense models. Dynamic Sparse Training achieves prior art performance compared with other sparse training algorithms on various network architectures. Additionally, we have several surprising observations that provide strong evidence to the effectiveness and efﬁciency of our algorithm. These observations reveal the underlying problems of traditional three-stage pruning algorithms and present the potential guidance provided by our algorithm to the design of more compact network architectures.

Details

YNIMG Journal 2007 Journal Article

Laminar profiles of functional activity in the human brain

David Ress
Gary H. Glover
Junjie Liu
Brian Wandell

Functional magnetic resonance imaging (fMRI) data were obtained in human visual cortex using sub-millimeter voxels at a field strength of 3 T. Reliable functional signals were largely confined to the gray matter and these responses measure the retinotopic organization of visual cortex. Functional signals were further characterized with respect to their laminar position within the cortical gray matter. The laminar response profiles during our visuospatial attention task, normalized for cortical thickness, had a stereotypical shape, with a peak in the superficial gray matter and declining in the deeper layers. The thickness of the sheet producing functional signals was in excellent agreement with the estimated structural thickness of the gray matter throughout early visual cortex (error<0. 5 mm). Thickness measurements were highly repeatable from session-to-session (error<0. 4 mm). Hence, it is feasible and useful to use high-resolution fMRI to measure laminar activity profiles. The ability to distinguish signals arising in different lamina has significant potential scientific and clinical applications.

Details DOI