The DOI matches its Crossref record; 47 of the 47 references it cites were found in a registry or an index.
ActivityNet-QA: A Dataset for Understanding Complex Web Videos via Question Answering
Open accessAbstract
Recent developments in modeling language and vision have been successfully applied to image question answering. It is both crucial and natural to extend this research direction to the video domain for video question answering (VideoQA). Compared to the image domain where large scale and fully annotated benchmark datasets exists, VideoQA datasets are limited to small scale and are automatically generated, etc. These limitations restrict their applicability in practice. Here we introduce ActivityNet-QA, a fully annotated and large scale VideoQA dataset. The dataset consists of 58,000 QA pairs on 5,800 complex web videos derived from the popular ActivityNet dataset. We present a statistical analysis of our ActivityNet-QA dataset and conduct extensive experiments on it by comparing existing VideoQA baselines. Moreover, we explore various video representation strategies to improve VideoQA performance, especially for long videos.
Overview
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“Overall, they obtained promising results on the TVQA [129] and ActivityNetQA [130] datasets.”
From Automatic question generation: a review of methodologies, datasets, evaluation metrics, and applications (2023) · cites this paper as [130]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Very Deep Convolutional Networks for Large-Scale Image Recognition
Simonyan, K., & Zisserman, A. (2014) · arXiv · doi:10.48550/arxiv.1409.1556 · Cited by 114,148 · cited 2×
“As we have obtained the sampled video unit set, for each unit ui, we used the VGG-16 network pre-trained on the ImageNet dataset (Simonyan and Zisserman 2014) to extract the appearance feature (the fc7 feature xi ∈ R(4096)) given the central frame of the unit, and the C3D network pretrained on the Sport-1M dataset (Tr…”
MethodHochreiter, S., & Schmidhuber, J. (1997) · Neural Computation · doi:10.1162/neco.1997.9.8.1735 · Cited by 102,318 · cited 2×
“E-VQA is the extension of an ImageQA baseline (Antol et al. 2015), where one long-short term memory (LSTM) network (Hochreiter and Schmidhuber 1997) is used to encode all words in the question and another different LSTM network is used to encode the frames in the video.”
BackgroundMethodGlove: Global Vectors for Word Representation
Pennington, J., Socher, R., & Manning, C. (2014) · Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.3115/v1/d14-1162 · Cited by 34,071 · cited 1×
“All models use the pre-trained 300-dimensional GloVe embedding (Pennington, Socher, and Manning 2014) to initialize the question embedding layer.”
MethodLearning Spatiotemporal Features with 3D Convolutional Networks
Tran, D., Bourdev, L., Fergus, R., Torresani, L., & Paluri, M. (2015) · 2015 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2015.510 · Cited by 9,593 · cited 2×
“As we have obtained the sampled video unit set, for each unit ui, we used the VGG-16 network pre-trained on the ImageNet dataset (Simonyan and Zisserman 2014) to extract the appearance feature (the fc7 feature xi ∈ R(4096)) given the central frame of the unit, and the C3D network pretrained on the Sport-1M dataset (Tr…”
MethodShow, Attend and Tell: Neural Image Caption Generation with Visual Attention
Xu, K., Ba, J., Kiros, R., Cho, K., Courville, A., Salakhutdinov, R., Zemel, R., & Bengio, Y. (2015) · arXiv · doi:10.48550/arxiv.1502.03044 · Cited by 8,248
Bottom-Up and Top-Down Attention for Image Captioning and Visual Question Answering
Anderson, P., He, X., Buehler, C., Teney, D., Johnson, M., Gould, S. J., & Zhang, L. (2018) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2018.00636 · Cited by 5,416 · cited 1×
“…shown recent success, with many approaches proposed to investigate the key components of this task, e.g. , discriminative feature representation (Anderson et al. 2018), multi-modal fusion (Kim et al. 2017a; Yu et al. 2017; Yu et al. 2018a) and visual reasoning (Nam, Ha, and Kim 2016; Lu et al.…”
BackgroundLong-term recurrent convolutional networks for visual recognition and description
Donahue, J., Hendricks, L. A., Guadarrama, S., Rohrbach, M., Venugopalan, S., Darrell, T., & Saenko, K. (2015) · 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2015.7298878 · Cited by 5,228 · cited 2×
“These advances stimulated research into bridging the semantic connections between vision and language, such as in visual captioning (Donahue et al. 2015; Xu et al. 2015), visual grounding (Rohrbach et al. 2016; Chen, Kovvuri, and Nevatia 2017; Yu et al. 2018b) and visual question answering…”
BackgroundVisual Genome: Connecting Language and Vision Using Crowdsourced Dense Image Annotations
Krishna, R., Zhu, Y., Groth, O., Johnson, J., Hata, K., Kravitz, J., Chen, S., Kalantidis, Y., Li, L.-J., Shamma, D. A., et al. (2017) · International Journal of Computer Vision · doi:10.1007/s11263-016-0981-7 · Cited by 5,107 · cited 3×
“datasets, such as Visual Genome (Krishna et al. 2016) and VQA (Antol et al.”
BackgroundMethodCIDEr: Consensus-based image description evaluation
Vedantam, R., Zitnick, C. L., & Parikh, D. (2015) · 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2015.7299087 · Cited by 4,845 · cited 1×
“We evaluate the average similarities of each two natural language sentences using CIDEr score (Vedantam, Lawrence Zitnick, and Parikh 2015), setting an empirical threshold to the average CIDEr score, and manually checking samples that did not reach the threshold.”
MethodVQA: Visual Question Answering
Antol, S., Agrawal, A., Lu, J., Mitchell, M., Batra, D., Zitnick, C. L., & Parikh, D. (2015) · 2015 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2015.279 · Cited by 4,487
Research works printed but not linked: 1
- Sukhbaatar, S.; Weston, J.; Fergus, R.; et al. 2015. End-toend memory networks. In NIPS , 2440-2448.
Cited by
Sorted by citations
Items 201–282 of 282
AssistSR: Affordance-centric Question-driven Video Segment RetrievalPreprint
Lei, S. W., Wang, Y., Mao, D., Gao, D., & Shou, M. Z. (2021) · arXiv (Cornell University) · Cited by 2
Spatiotemporal fusion personality prediction based on visual information
Xu, J. D., Tian, W., Lv, G., & Fan, Y. (2023) · Multimedia Tools and Applications · doi:10.1007/s11042-023-15537-0 · Cited by 1
“Small well-labeled datasets such as KTH, Hollywood2, Weizmann [3]; medium-sized datasets such as UCF101, Thumos’14 and HMDB51, which exceed 50 behavioral categories; large datasets such as Sports-1M, YFCC-100M , FCVID dataset, ActivityNet dataset, YouTube-8M, etc. [27].”
BackgroundSemantic Distance-Aware Cross-Modal Attention Mechanism for Video Question Answering
Liu, M., Zhou, F., Wang, R., Zhao, B., & Zhang, F. (2025) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2025.3571119 · Cited by 3
End-to-End 3D point cloud question answering via state space models
Qorsham, H., El-Deeb, R., & Essa, E. (2025) · Applied Soft Computing · doi:10.1016/j.asoc.2025.113404 · Cited by 1
SeriesBench: A Benchmark for Narrative-Driven Drama Series Understanding
Zhang, C., Lei, Y., Liu, Z., Leng, H., Liu, S., Gao, T., Liu, Q., & Wang, Y. (2025) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52734.2025.02700 · Cited by 1
Luo, Z., Wu, H., Li, D., Ma, J., Kankanhalli, M., & Li, J. (2025) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52734.2025.00792 · Cited by 6
“To address these limitations, benchmarks like ActivityNet-QA [68] and EgoSchema [49] have extended video durations to approximately 180 seconds on average.”
VisiQuest:Video-Question Answering with Advanced Vision-Language AI
Praneeth, S., Kumar, R. P., & Pallavi, G. (2025) · International Conference on Computing Technologies (ICOCT) · doi:10.1109/icoct64433.2025.11118375 · Cited by 1
“To bring this limitation into lime light the ActivityNet-QA [13] dataset was introduced, offering fully annotated, long-duration videos with diverse question types, enabling improved VideoQA research.”
Wang, H., Lai, C., & Ge, W. (2025) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2025.3607780 · Cited by 2
Hu, Y., Yang, Z., Wang, S., Qian, S., Wen, B., Yang, F., Gao, T., & Xu, C. (2025) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3746027.3758311 · Cited by 1
“Traditional video question answering datasets, such as [6] [20] [25], typically require annotators to watch videos, extract the core content, and generate corresponding question-answer pairs based on predefined standards.”
FineQuest: Adaptive Knowledge-Assisted Sports Video Understanding via Agent-of-Thoughts Reasoning
Haodong, C., Huang, H., Yin, X., & Shao, D. (2025) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3746027.3754885 · Cited by 1
Parse, Align and Aggregate: Graph-Driven Compositional Reasoning for Video Question Answering
Li, J., Liao, Z., Xiao, F., Li, T., Zhang, Q., Zhao, H., Niu, L., Chen, G., Zhang, L., & Jiang, C. (2026) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2026.3650864 · Cited by 1
“We evaluate the performance of our QPVA 3 framework on 6 VideoQA benchmarks, including our QPVA 3 Bench, AGQA-Decomp [11], ActivityNet-QA [51], Causal-VidQA [5], STAR [6], and MVBench [9].”
Zhou, H., Peng, X., Kendre, S., Ryoo, M. S., Savarese, S., Xiong, C., & Niebles, J. C. (2025) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw69036.2025.00451 · Cited by 1
LoVR: A Benchmark for Long Video Retrieval in Multimodal Contexts
Liang, H., Cai, Q., Han, Z., Dong, H., Qiang, M., An, R., Xu, Q., Cui, B., & Zhang, W. (2026) · International World Wide Web Conference (WWW) · doi:10.1145/3774904.3792099 · Cited by 17
“Notably, long-form videos, such as lectures, documentaries, and live streams lasting from minutes to hours, constitute a significant and growing portion of this information ecosystem.”
VideoAds for Fast-Paced Video Understanding
Zhang, Z., Dou, W., Peng, L., Pan, H., Bagci, U., & Gong, B. (2025) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51701.2025.02025 · Cited by 1
“For example, ActivityNet-QA [44] focuses on human activity recognition, and MSVD-QA [41] are generated by”
One Trajectory, One Token: Grounded Video Tokenization Via Panoptic Sub-Object Trajectory
Zheng, C., Zhang, J., Salehi, M., Gao, Z., Iyengar, V., Kobori, N., Kong, Q., & Krishna, R. (2025) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51701.2025.02150 · Cited by 4
“We train both VideoLLMs for 3 epochs with a batch size of 32 and evaluate the models on various QA benchmarks, including VideoMME [16], NextQA [63], ActivityNetQA [67], MLVU [73], TempCompass [37], and MovieChat [52].”
ST-Think: How Multimodal Large Language Models Reason About 4D Worlds from Ego-Centric Videos
Wu, P., Liu, Y., Liu, M., & Shen, J. (2026) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv61042.2026.00502 · Cited by 18
Beyond the Highlights: Video Retrieval with Salient and Surrounding Contexts
Bang, J., Moon, Y.-B., Oh, T.-H., & Joo, K. (2026) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv61042.2026.00217 · Cited by 1
“These datasets cover a range of tasks, including video classification [4, 5, 16, 19], captioning and retrieval [2, 3, 7, 8, 25, 29, 36, 40], as well as question answering [15, 34, 38].”
Accelerating Multimodal Large Language Models with Prior-Corrected Token Reduction
Chen, Z., Cai, Y., Guo, J., Cai, T., Yin, J., & Chen, Z. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-36839-3_13 · Cited by 1
Adaptive Proposal on Spatio-Temporal Graph in Audio-Visual Question Answering
Xie, Z., Yang, Y., Yu, Y., Wang, J., Peng, B., Li, X., & Liu, Y. (2026) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2026.3714320 · Cited by 1
“Dataset: The proposed method is evaluated on four widely-used datasets: MUSIC-AVQA [18], MUSIC-AVQA-R [69], AVQA [20], and ActivityNet-QA [70].”
Shen, Y., & Unberath, M. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-37174-4_11 · Cited by 1
Classical Visual Question Answering
Wu, Q., Wang, P., Wang, X. Y., He, X., & Zhu, W. (2022) · Advances in computer vision and pattern recognition · doi:10.1007/978-981-19-0964-1_4 · Cited by 1
Exploring Visual Multiple-Choice Question Answering with Pre-trained Vision-Language Models
Tran, G.-N., Luu, D.-T., & Thin, D. V. (2025) · Lecture notes in computer science · doi:10.1007/978-981-96-2641-0_22 · Cited by 1
“Furthermore, VQA research has extended the task to different variations of visual input, including videos [48,52], scene text [6], and documents [35,42], broadening the applicability of the task across different domains.”
MELA: Multi-Event Localization Answering Framework for Video Question Answering
Han, J. G., Kim, S., & Park, H. (2025) · ACM Symposium on Applied Computing · doi:10.1145/3672608.3707973 · Cited by 1
“Early VideoQA benchmarks [29, 30, 34, 35] primarily focused on questions regarding the actions performed by characters in a video or the attributes of objects featured within it.”
How Good is my Video-LMM? Complex Video Reasoning and Robustness Evaluation Suite for Video-LMMs
Khattak, M. U., Naeem, M. F., Hassan, J., Naseer, M., Tombari, F., Khan, F. S., & Khan, S. H. (2025) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw67362.2025.00349 · Cited by 1
“2, both Gemini and GPT variants improve over open-source models and achieve high gains across all evaluation dimensions.”
MECD+: Unlocking Event-Level Causal Graph Discovery for Video Reasoning
Chen, T., Liu, H., Wang, Y., Chen, Y., He, T., Gan, C., He, H., & Lin, W. (2025) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2025.3626549 · Cited by 2
Xu, L., Huang, H., & Liu, J. (2021) · arXiv (Cornell University) · Cited by 1
Pawar, K., Halbe, S., Patil, V., Gupta, V., & Attar, V. (2024) · SSRN Electronic Journal · doi:10.2139/ssrn.4880715 · Cited by 1
iLLaVA: An Image is Worth Fewer Than 1/3 Input Tokensin Large Multimodal ModelsPreprint
Hu, L., Shang, F., Wan, L., & Feng, W. (2024) · Qeios · doi:10.32388/69kflk · Cited by 1
Wu, Q., Wang, P., Wang, X. Y., He, X., & Zhu, W. (2022) · Advances in computer vision and pattern recognition · doi:10.1007/978-981-19-0964-1_8
A Framework for Video-Text Retrieval with Noisy Supervision
Vaseqi, Z., Fan, P., Clark, J. J., & Levine, M. D. (2022) · INTERNATIONAL CONFERENCE ON MULTIMODAL INTERACTION · doi:10.1145/3536221.3556597
“Compared to other video-text datasets such as ActivityNet [37] and THUMOS [15], in our Hockey dataset the diference between action classes is more nuanced: this emphasizes the need for learning more discriminative features rather than principally relying on high level features, such as the appearance features derived…”
BackgroundMethodInfluentialVideo-and-Language (VidL) models and their cognitive relevance
Zonneveld, A. W., Albert Gatt, A., & Calixto, I. (2023) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw60793.2023.00040
“VidQA provides fine-grained visual understanding [142] and in a typical automatic evaluation scenario one compares the model’s generated output against one or more ground truth answers [4].”
BackgroundLuo, Y., Wang, R., Zhang, F., Zhou, F., Liu, M., & Feng, J. (2024) · Neural Computing and Applications · doi:10.1007/s00521-024-09482-8
LogSay: An Efficient Comprehension System for Log Numerical Reasoning
Qi, J., Luan, Z., Huang, S., Fung, C., & Yang, H. (2024) · IEEE Transactions on Computers · doi:10.1109/tc.2024.3386068 · Cited by 1
“ActivityNet-QA [12] Video ♣ Medium LifeQA [13] Video ♣ Medium VQuAD [7] Video ♣ Hard SQuAD [1] Wikipedia ♣ Simple HotpotQA [14] Wikipedia ♠♣ Medium Drop [2] Wikipedia ♣ Medium 2WikiMultiHopQA [15] Wikipedia ♠♣ Hard NewsQA [16] News ♣ Simple JEC-QA [8] Legal ♣ Hard Event-QA [9] Knowledge Graphs ♠♣…”
Leveraging LLM’s Mastery in Video Question Answering
Rakshitha, R. T., Desai, P., Sujatha, C., & Gangisetty, S. (2025) · Lecture notes in networks and systems · doi:10.1007/978-981-96-2700-4_38
LBA: Multi-Scale Video Segment Sampling for Open-Ended Video Question Answering
Wang, J., & Han, Y. (2025) · Lecture notes in computer science · doi:10.1007/978-981-96-9856-1_9
Watanabe, K., Masuda, S., Xiao, L., & Yamasaki, T. (2025) · International Conference on Automatic Face and Gesture Recognition · doi:10.1109/fg61629.2025.11099276
“Some VQA benchmarks [14], [9], [13], [28] include action recognition tasks in which individuals are identified by their physical characteristics.”
SRIN: Structured Reasoning Integration Network for Robust Video Question AnsweringPreprint
Yamada, K. (2025) · Preprints.org · doi:10.20944/preprints202508.1438.v1
APSam: An Aggregating-Then-Pruning Sampler for Question-Conditional Denoising
Liang, J., Jiang, S., Tang, W., Wang, N., Wang, Z., Mao, X., Lv, K., Liu, M., & Qin, B. (2025) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2025.3609803
“…is minimal and the added latency is very low. e) Performance on Long-form Video: To further verify and analyze APSam ’s capability in long-form understanding, we evaluate it on three mainstream video understanding benchmarks, including LongVideoBench [60], VideoMME [61], and Activitynet-QA [62].”
Exploring the Application of Visual Question Answering (VQA) for Classroom Activity Monitoring
Vu, S. T., Trung, H. P., Nguyen, D. M., Hoang, H. M., Le, N. H., Pham, T. H., & Mai, T. T. (2025) · doi:10.1145/3746274.3760394
“However, none of the existing datasets reflect the specific context of classroom environments , which involves complex learning behaviors and interactions between teachers and students.”
Jiang, P., Li, H., Zhao, L., Chao, F., Yan, K., Ding, S., & Ji, R. (2025) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3746027.3755792 · Cited by 4
“Finally, we further verify our method through visualization results and ablation study.”
Ddog: optimizing multi-hop inference via dual-driven retrieval and reasoning path
Chen, Y., Gu, B., Gao, L., Fu, K., Qu, Y., & Cui, L. (2025) · Machine Learning · doi:10.1007/s10994-025-06917-8
“(3) In multi-hop path break scenarios (queries with more than 30% missing data in the Freebase graph in CWQ), DDOG’s F1 score improves by 13.6% over GraphRAG, demonstrating the effectiveness of dynamic knowledge expansion in mitigating path break issues.”
InfluentialTemporalCook: Benchmarking Temporal and Procedural Reasoning in Multimodal Large Language Models
Zarei, M. R., Akkasi, A., Hassan, M., & Komeili, M. (2025) · doi:10.1145/3746276.3760469
Beyond Description: Cognitively Benchmarking Fine-Grained Action for Embodied Agents
Liu, D., Xu, C., Chen, W., Zhang, S., Wang, J., Deng, J., Sun, B., & Liu, Y. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-37010-5_21 · Cited by 1
Momentor++: Advancing Video Large Language Models With Fine-Grained Long Video Reasoning
Li, J., Gao, M., He, X., Tang, S., Zheng, W., Xiao, J., Wang, M., Chua, T., & Zhuang, Y. (2026) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2026.3656169
“As shown in Table III, Momentor and Momentor++ consistently outperform their backbone counterparts across MSVD-QA [59], MSRVTT-QA [59], ActivityNet-QA [60], NExT-QA [61], and YouCook2 [62].”
Self-citationAdversarial Video Promotion Against Text-to-Video Retrieval
Tian, Q., Lin, C., Zhao, Z., Liu, S., Li, Q., & Shen, C. Q. (2026) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2026.3657094 · Cited by 1
“For dataset-wise evaluations, we test on three prevailing datasets using the Singularity model—ActivityNet [48], DiDeMo [14], and MSR-VTT.”
InfluentialSelf-citationBeyond Real Versus Fake: Towards Intent-Aware Video Analysis
Atreya, S., Quignon, N., Chopin, B., Das, A., & Dantcheva, A. (2026) · IEEE Transactions on Biometrics Behavior and Identity Science · doi:10.1109/tbiom.2026.3663642 · Cited by 1
“Progress in video understanding is fueled by increased volumes of training datasets [25], [26], [27], [28], [29].”
Sun, G., Liang, L., Qiu, J., Zhang, L., & Lang, F. (2026) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2026.3665004
“Recent VideoQA datasets, including TVQA+ [22] and ActivityNet-QA [21], highlight that state-of-the-art models struggle to answer compositional questions due to various sources of error.”
SELongVLM: Empowering Long Video Language Models With Self-Corrective Clip Selection
Zhang, K., Yang, Z., Han, M., Zhuge, Y., Hao, H., Li, C., Li, Z., & Chang, X. (2026) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2026.3673141
“They exceed random performance levels on various types of video understanding benchmarks by leveraging broad and diverse training datasets, ranging from seconds-long video interactions such as [29], [31] to minutes-long videos like [15], [32].”
PerceptionComp: A Video Benchmark for Complex Perception-Centric Reasoning
Li, S., Zhao, Z., Deng, H., Ma, Z., Tian, S., Liu, Z., Hu, Y., Wu, H., Dong, Y., Liu, B., et al. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-37553-7_14
Self-citationWorst-Projection-Gap Guided Pruning for Long-Tail Visual Token CoveragePreprint
Liu, Y., Su, Y., Hao, W., & Wang, L. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6505413
Video question answering with limited supervision
Engin, D. (2024) · theses.fr (ABES) · doi:10.70675/c9c7dcf2z4ca3z4e91z9d27zdacd00f76581
Multimodal deep learning for audiovisual production
Ouenniche, K. (2023) · theses.fr (ABES) · doi:10.70675/a01becf4z4ed9z4955z940cz49c821c1b8a8
Ethical and Practical Trade-Offs in Human Action Recognition Models for Assistive Systems
Chiroiu-Jitianu, D., Nan, M., & Mocanu, I. G. (2025) · International Symposium on Symbolic and Numeric Algorithms for Scientific Computing (SYNASC) · doi:10.1109/synasc69064.2025.00068
“Evaluation was conducted on widely recognized video question-answering datasets: MSVD-QA [23], ActivityNet-QA [24], MSRVTT-QA [25] and TGIF-QA [26].”
Zhang, Y., Li, H., Liu, J., & Yue, X. (2025) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51701.2025.00132
“…[59], Flickr [75], ClothoV1 [19], ClothoV2 [19], Audio-Caps [41], MSRVTT [98], YouCook2 [115], VALOR-32K [9], VATEX [91], DEDeMo [4], and ANET [106]), 7 caption tasks (COCO, ClothoV1, ClothoV2, AudioCaps, MSRVTT, YouCook2, VALOR-32K), and 6 QA tasks (TGIF [56], MSVD [96], VQAv2 [28], MSRVTT,…”
Learning Beyond Still Frames: Scaling Vision-Language Models with Video
Zhang, Y., Jing, H., Liu, J., & Yue, X. (2025) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51701.2025.02082
“…achieves new state-of-the-art performance on leading vision-language benchmarks , as shown in Figure 1, including MME [21], MMBench [47], and MMVet [77], among others, as well as on video understanding benchmarks such as EgoSchema [51], VideoMME [22], Activi-tyNet [78], and other benchmarks.”
ViSpeak: Visual Instruction Feedback in Streaming Videos
Fu, S., Yang, Q., Li, Y.-M., Peng, Y.-X., Lin, K.-Y., Wei, X., Hu, J.-F., Xie, X., & Zheng, W. (2025) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51701.2025.02022 · Cited by 21
Bringing RNNs Back to Efficient Open-Ended Video Understanding
Xu, W., Song, E., Chai, W., Wen, X., Tian, Y., & Wang, G. (2025) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51701.2025.02177
“Token Order ANet [102] VATEX [89] VDC [11] sequential input token order, which is disrupted by the visual token merging process.”
Chen, Y., Wang, J., Zhang, Z., Yi, J., Zhang, X., Zou, Y., Cai, Z., Yuan, J., Li, X., Yang, H., et al. (2026) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv61042.2026.00413
Shin, J., Hur, C., Cho, S., Choi, J., & Park, H. (2026) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv61042.2026.00738
“We demonstrate the effectiveness of RefineVQA by comparing it with existing approaches on the NExT-QA [34], EgoSchema [23], Activity-Net [40], Video-MME [6], and StreamingBench [17] benchmarks, as shown in Tab.”
SVI-Bench: A Dynamic Microworld for Strategic Video Intelligence
Pan, Y., Han, Y., Ha, S., Islam, M. M., Zhang, B., Torresani, L., & Bertasius, G. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-37095-2_12
Wu, Z., Liu, Z., Chen, A., Zhang, J., Li, R., Ge, H., Wang, Z., Xiao, C., & Liang, C. (2026) · International Journal of Computer Vision · doi:10.1007/s11263-026-02898-w
“Early Factoid VideoQA datas ets (Xu et al., 2017; Yu et al., 2018, 2019; Li et al., 2020; Liu et al., 2024; Li et al., 2024; Wu et al., 2024; Mangalam et al., 2023) primarily focus on simple visual facts within short-range video clips, such as object recognition (Li et al., 2020; Xiao et al., 2021;…”
InfluentialReflective Cross-Granularity Grounding with Preference Optimization for Long Video Understanding
Feng, W., Wang, X. Y., Chen, H., Zhan, Y.-W., Song, Z., Huang, B., Zheng, K., & Zhu, W. (2026) · International Conference on Multimedia Retrieval · doi:10.1145/3805622.3810717
Spectral Evolution-Guided Token Pruning in Multimodal Large Language Models
Chen, B., Cai, Y., Luo, Y., Zhang, Y., Yin, J., & Chen, Z. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-37517-9_13
REAP: Relevance-Guided Evidence-Unit Keyframe Extraction for Long Video Question Answering
Yang, M., Qu, H., & Zhao, J. Z. (2026) · International Symposium on Computer Technology and Information Science (ISCTIS) · doi:10.1109/isctis70043.2026.11572599
VisReflect: Latent Visual Reflection for Fine-Grained Perception in Long Visual Context
Shen, X., & Elhoseiny, M. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-37038-9_36
LongEgoRefer: A Benchmark for Long-Form Egocentric Video Referring Expression Comprehension
Kato, S., Miyanishi, T., Kurita, S., Ukai, M., Inoue, N., & Chu, C. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-37261-1_18
GuideMe: Multi-domain Task Guidance and Intervention in Streaming Video
Liu, F., Chen, J., Xu, K., Liu, Y., Guan, H., Lu, X., Yang, B., Hancke, G. P., Liu, R., & Lau, R. W. H. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-37592-6_4
SparseVLM+: Visual Token Sparsification With Improved Text-Visual Attention Pattern
Zhang, Y., Ma, J., Zhang, Q., Fan, C.-K., Zheng, W., Cheng, K., Lu, J., & Zhang, S. (2026) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2026.3718245
“We validate Sparse-VLM+ on eight common video understanding benchmarks, including TGIF-QA [72], MSVD-QA [73], MSRVTT-QA [73], ActivityNet-QA [74], TempCompass [75], LongVideoBench [76], MLVU [77], and VideoMME [78], where video-question pairs and video lengths are highly disproportionate.”
EchoVision: A Decoupled Audio-Visual Retrieval Framework for Video Question Answering
Sarkar, S., Chakraborty, J., Nandi, S., Samui, S., & Nandi, S. (2026) · International Conference on Signal Processing and Communications (SPCOM) · doi:10.1109/spcom71105.2026.11622962
“1-mini ( sys ∥ ctx ∥ q ) 10: return a ∗ ActivityNet-QA (ACNT) [9]: real-world activities in diverse settings; 86 questions (mainly PD, RS, ST) from 7 videos ( ∼ 3.3 min avg).”
InfluentialScene Graph for Vietnamese Video Understanding: An Agentic Approach with Reasoning
Vo, Q. N. T., Phan, D. T., & Quan, T. T. (2026) · Communications in computer and information science · doi:10.1007/978-981-92-2600-9_12
Long Story Short: Story-level Video Understanding from 20K Short Films
Ghermi, R., Wang, X., Kalogeiton, V. S., & Laptev, I. (2026) · International Journal of Computer Vision · doi:10.1007/s11263-026-02940-x · Cited by 22
“…datasets have been made available through the years, covering visual descriptions [9, 76], temporal action reasoning [71], compositional reasoning [22], social intelligence [82], instructional videos [77], egocentric videos [27, 37, 41], movies [31, 40, 61], among others [26, 26, 34, 69, 74, 81].”
Towards general auditory intelligence for machine listening and speaking
Wang, S., Jin, Z., Tang, C., Li, Q., Li, B., Chen, C., Hu, Y., Yu, W., Li, Y., Zhuang, J., et al. (2026) · Nature Machine Intelligence · doi:10.1038/s42256-026-01281-1 · Cited by 6
Ask, Then Perceive: Question-Guided Cross-Modal Perception for Audio-Visual Question Answering
Ye, Q., Wang, J., Chen, C., Yu, S., Ma, H., Lin, X., & Yu, Z. (2026) · International Conference on Bioinformatics and Intelligent Computing (BIC) · doi:10.1145/3809986.3810022
Self-citationTo Keep or Not to Keep: Learning KV Cache Retention in Disaggregated LLM Serving Systems
Liu, D., Yu, Y., Jiang, E., Wang, S., & Wu, Y. N. (2026) · ACM International Systems and Storage Conference (SYSTOR) · doi:10.1145/3793230.3837769
Training-Free Event-Aware Memory Enhancement for Streaming Video Question Answering
Song, Y. (2026) · International Conference on Image, Vision and Computing (ICIVC) · doi:10.1109/icivc70586.2026.11686687
“Early methods relied on joint visual-textual embeddings with attention mechanisms to localize relevant frames [10], [11].”
Li, Y., Yang, Y., Yu, F., Hu, J., Fang, Z., & Huang, L. (2026) · International Conference on Big Data and Information Analytics (BigDIA) · doi:10.1109/bigdia71496.2026.11684568
“These methods typically rely on manual annotations [4] or rule-based designs [5], with evaluation centering on the model’s ability to align visual content with language queries.”
MotionAtlas: A High-Quality Dataset and Benchmark for Dense Motion Captioning
Liu, W., Wang, H., Gao, K., Wang, Y., Zhou, Y., Ren, Z., Mai, J., Wang, A., Li, Y., Li, J., et al. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-37467-7_15 · Cited by 1
VSI: Visual–Subtitle Integration for Keyframe Selection to Enhance Long Video Understanding
He, J., 盛, 红., Li, J., Guo, W., Hu, X., & Xiong, H. (2026) · doi:10.1109/cvprf71909.2026.00864
“Long-form video analysis [26, 47] poses greater and more persistent challenges than short-video or image tasks [48], primarily due to inherently complex temporal dynamics and substantial content redundancy [14, 31, 51, 52].”
HippoMM: Hippocampal-Inspired Multimodal Memory for Long Audiovisual Event Understanding
Lin, Y., Zhang, J., Wang, Q., Ye, H., Fu, Y., Liu, Y., Li, H. H., & Chen, Y. (2026) · 2026 IEEE/CVF Conference on Computer Vision and Pattern Recognition- FINDINGS Track (CVPRF) · doi:10.1109/cvprf71909.2026.00573 · Cited by 1
“Unlike existing benchmarks that focus on sequential comprehension [22] or simple retrieval tasks [39], HippoVlog evaluates episodic memory functions through partial cue reconstruction and semantic consolidation questions.”
One Identity, Many Roles: Multimodal Entity Coreference for Enhanced Video Situation Recognition
Darur, B., Garg, A., & Tapaswi, M. (2026) · doi:10.1109/cvprf71909.2026.00794
“…[15, 21, 55, 83], spatio-temporal action detection [17, 58], and visual tracking [4, 11, 12, 47, 74]; and (ii) video-language tasks such as dense video captioning [22, 72, 84], video question answering [13, 20, 29, 59, 77], referring expression grounding and segmentation [24, 32, 41, 50, 71].”
Taming Hallucinations: Boosting MLLMs' Video Understanding via Counterfactual Video Generation
Huang, Z., Wen, H., Hao, A., Song, B., Wu, M., Wu, J., Chu, X., Lu, S., & Wang, H. (2026) · doi:10.1109/cvprf71909.2026.00783
Nguyen, L. T., Yu, Z., Hang, S. L. Y., An, S., Lee, J., Ban, Y., Chung, S., Nguyen, T.-H., Maeng, J., Lee, S., et al. (2026) · doi:10.1109/cvprf71909.2026.00222
Self-citation
Metadata
- Authors
- Yu, Z., Xu, D., Yu, J., Yu, T., Zhao, Z., Zhuang, Y., & Tao, D.
- Affiliations
- Zhejiang University (CN)
- Hangzhou Dianzi University (CN)
- The University of Sydney (AU)
- Venue
- Proceedings of the AAAI Conference on Artificial Intelligence
- Published
- 2019-07-17
- Volume
- 33
- Issue
- 01
- Pages
- 9127-9134
- ISSN
- 2159-5399 (ISSN-L)
- 2374-3468
- OpenAlex
- W2964220823
- Cited by
- 281
- References
- 43
- Citations by year
- 2026: 44 · 2025: 73 · 2024: 47 · 2023: 39 · 2022: 26 · 2021: 32 · 2020: 15 · 2019: 5
- FWCI
- 6.36
- Type
- conference paper
- Language
- en
- Open access
- diamond
- doi.org
- Licence
- www.aaai.org/ (unspecified)
- Locations
- Proceedings of the AAAI Conference on Artificial Intelligence (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- AAAI Publications (The Association for the Advancement of Artificial Intelligence (AAAI)) (submittedVersion)
- arXiv (Cornell University)
- Funders
- National Natural Science Foundation of China (61602405, Grant 61622205, 61622205, 61702143, 61836002)
- China Knowledge Centre for Engineering Sciences and Technology
- Australian Research Council (FL170100117, IH180100002, DP180103424)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Multimodal Machine Learning Applications
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Human Pose and Action Recognition
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Video Analysis and Summarization
- Keywords
- video question answering, annotated dataset, online video, ActivityNet, video representation
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
