Video as Conditional Graph Hierarchy for Multi-Granular Question Answering

Authors: Junbin Xiao, Angela Yao, Zhiyuan Liu, Yicong Li, Wei Ji, Tat-Seng Chua2804-2812

AAAI 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Despite the simplicity, our extensive experiments demonstrate the superiority of such conditional hierarchical graph architecture, with clear performance improvements over prior methods and also better generalization across different type of questions.
Researcher Affiliation Academia Department of Computer Science, National University of Singapore
Pseudocode No The paper describes the model architecture and operations, but it does not present them in a structured pseudocode or algorithm block.
Open Source Code No The paper does not explicitly state that source code for the methodology is openly available or provide a link to a code repository.
Open Datasets Yes We experiment on four Video QA datasets that challenge the various aspects of video understanding: TGIF-QA (Jang et al. 2019), MSRVTT-QA and MSVD-QA, NEx T-QA (Xiao et al. 2021).
Dataset Splits No The paper mentions using "validation sets" (e.g., "We analyze our model on the validation sets of NEx T-QA and MSRVTT-QA") but does not specify the exact split percentages, sample counts, or detailed methodology for creating these splits in the main text.
Hardware Specification No The paper does not provide specific hardware details such as GPU/CPU models, memory, or detailed computer specifications used for running the experiments.
Software Dependencies No The paper mentions various software components and models like 'BERT model', '3D version Res Ne Xt-101', 'Res Net-101', 'Bi-GRU', and 'Adam optimizer', but it does not specify their exact version numbers.
Experiment Setup Yes For training, we adopt a two-stage scheme by firstly training the model with learning rate lr = 10 4 and then fine-tune the best model obtained in the 1st stage with a smaller lr, e.g., 5 10 5. For both stages, we train the models by using Adam optimizer with batch size of 64 and maximum epoch of 25. The dimension of the models hidden states is d = 512 and the default number of graph layers in QGA is H = 2.