About MeI am currently a first-year PhD student in computer science at UCLA, advised by Prof. Song-Chun Zhu. Previously, I was an undergraduate in the Department of Electrical Engineering in Shanghai Jiao Tong University, advised by Prof. Xinbing Wang and Prof. Weinan Zhang.
I’ve been drawing inspiration from cognitive science, and endowing machines with human-like concept learning abilities via neural-graphical-symbolic learning. Specifically:
Neural: neural networks for natural language processing and computer vision.
Graphical: graph parsing via grammar learning.
Symbolic: symbolic reasoning, including program synthesis, inductive logic programming and theorem proving.
Publication & PatentYining Hong; Qing Li; Daniel Ciao; Siyuan Huang; Song-Chun Zhu, "Learning by Fixing: Solving Math Word Problems with Weak Supervision", AAAI-21
Yining Hong; Qing Li; Ran Gong; Daniel Ciao; Siyuan Huang; Song-Chun Zhu, "SMART: A Situation Model for Algebra Story Problems via Attributed Grammar", AAAI-21
Qing Li; Siyuan Huang; Yining Hong; Song-Chun Zhu, "A Competence-aware Curriculum for Visual Concepts Learning via Question Answering", ECCV2020 (Oral, 2%)
Qing Li; Siyuan Huang; Yining Hong; Yixin Chen; Yingnian Wu; Song-Chun Zhu, "Closed Loop Neural-Symbolic Learning via Integrating Neural Perception, Grammar Parsing, and Symbolic Reasoning", ICML2020
Yining Hong; Jialu Wang; Yuting Jia; Weinan Zhang; Xinbing Wang, “Academic Reader: An Interactive Question Answering System on Academic Literatures”, Proceedings of the Thirty-Third AAAI Conference on Artificial Intelligence, 2019
Jialu Wang; Yining Hong; Runqing Zhou; Xinbing Wang, “Learning to Read Academic Literature”, (manuscript), 2019.
MusicFormer president of SJTU Piano Association, playing the piano with love since 2001. Top 30 in Shanghai College Students' Piano Talent Contest. Jazz, Symphony and Musical Lover. Skilled at works of Liszt and Chopin. Held the first Piano Duo Concert in SJTU. Also can play a Chinese instrument called Pipa.
ResearchResearch Intern, Tencent AI Lab, advised by Piji Li.
Incorporating Hierarchical Graph Reasoning into Multi-turn Dialogue Generation
Research Assistant, IIoT Research Center, co-advised by Prof.Weinan Zhang and Prof.Xinbing Wang.
Learning to Read Academic Literature
Collected PaperQA, a machine reading comprehension dataset on academic papers, which consists of over 12,000 question-answer pairs on a set of over 1,800 academic abstracts via crowdsourcing.
Designed a new model which utilizes the shared query aware context representation as the base of sentence ranking and answer extraction.
Presented Academic Reader, an interactive system which leverages machine reading comprehension techniques to read academic literature and answer the relevant question for researchers.
Research Assistant, IIoT Research Center, advised by Prof.Xinbing Wang.
Acemap: An Academic Visualization System
Designed and implemented an academic search system to visualize the academic networks.
Applied academic maps to integrate the present academic data and crawled large data to obtain the database including 1.27 hundred million papers, 1.15 hundred million scholars and 24000+ journals.
Implemented detailed front-end and back-end functions of Acemap website. Analyzed papers, authors and affiliations in latest conferences including KDD, MOBICOM, OSDI, etc., to benefit more researchers.
Text classification Kaggle Competition
Ranked 1st in public leaderboard, 2nd in private. Best Presentation Award in the poster session.
Trained classification models to predict the categories of Chinese texts. Applied Text-CNN, Xgboost, Logistic Regression, Random Forest, Naive Bayes as basic models.
Implemented some ensembling above these models using techniques such as parameter disturbances, 1/rank weighted mean and rank weighted mean.
Code | Report | Poster
Link Prediction Kaggle Competition
Applied, compared and ensembled two models: TransE and Node2Vec to analyze the possible attributes of entities to infer missing links in an observed academic knowledge graph.
Classify CIFAR-10 $\&$ MNIST Datasats through Different Classifiers
Ranked 1st among 163 undergraduates.
Introduced and implemented different machine learning classifiers: KNN, Linear SVM, Kernel SVM, Fisher’s Linear Discriminant and Kernel Fisher Discriminant on CIFAR-10 and MNIST datasets.
Compared the advantages and disadvantages with regard to accuracy, easiness to implement, training and testing time, memory requirement and etc.
Code | Report
Comparisons for Multilayer Perceptron on MNIST Dataset
Ranked 1st among 163 undergraduates.
Implemented a neural network with one hidden layer.
Compared the training results using different activation functions, optimization algorithms, hyperparameters and gradient descent techniques.
Visualized the results with regard to MSE, classification accuracy.
File Transfer Application Based on Wi-Fi Direct
Developed a file transfer application based on Wi-Fi Direct to implement the file sharing between Android and PC terminal without wireless network.
Implemented the bidirectional send-and-receive function through socket communication and extended the group sending function at Android terminal.
Applied C# to realize the handshake and file transfer function at PC terminal, and also, between PC terminal and Android terminal.
Code for Android