Pith. sign in

REVIEW 4 major objections 5 minor 6 cited by

EgoExoBench: A Benchmark for First- and Third-person View Video Understanding in MLLMs

T0 review · 4 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The paper introduces EgoExoBench, a 7,330-question benchmark showing that current multimodal large language models fall roughly 42 points behind deliberate human accuracy when asked to reason across first- and third-person video views.

desk verdict A genuinely useful new benchmark, but the headline cross-view-reasoning claim is not yet fully supported until the numeric inconsistencies are fixed and single-view solvability is controlled. read the letter →

arxiv 2507.18342 v1 pith:K6L7Y2LH submitted 2025-07-24 cs.CV

Yuping He , Yifei Huang , Guo Chen , Baoqi Pei , Jilan Xu , Tong Lu , Jiangmiao Pang This is my paper · ORCID
classification cs.CV
keywords egocentricvideounderstandingexocentricmultimodallargelanguagemodelscross-viewreasoningquestionansweringbenchmarkconstructiontemporalspatialcorrespondence
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper introduces EgoExoBench, a benchmark of 7,330 four-choice video questions built from six paired egocentric-exocentric datasets. Its aim is to test whether multimodal large language models can do what humans do routinely: align the same action, object, person, or task across first- and third-person views, translate spatial information between those views, and reason about the order of events across two streams. The paper reports that the best models reach only 48 to 52 percent accuracy while humans deliberately answering the same questions reach 90.1 percent, and that egocentric-only pretraining and chain-of-thought prompting do not close the gap. The central claim is that cross-view reasoning is a distinct, currently missing capability that single-view video benchmarks do not measure.

What carries the argument

The carrying mechanism is the benchmark's construction pipeline rather than a single algorithm. EgoExoBench aggregates synchronized and asynchronous ego-exo video pairs from six public datasets; questions are produced by three strategies (annotation-derived templates, LLM generation with DeepSeek-V3, and human annotation for fine-grained spatial understanding); and a consistency check plus a vision-grounded filtering step together use Qwen2.5-32B to discard any question answerable from text alone. Each multiple-choice item presents one query video plus four candidate videos from the other viewpoint, with negative options chosen to share scene, action, or temporal cues so that only genuine cross-view reasoning selects the correct answer. This filtering is what lets the authors attribute low model scores to a missing cross-view capability rather than to language shortcuts.

What would settle it

Give an LLM the text-only version of a random sample of EgoExoBench questions (question and options, with no video frames) under the same standardized answer-extraction protocol; if accuracy on that set rises substantially above the 25 percent random baseline, the vision-grounded filter has leaked text-solvable items and the reported gap is inflated.

Watch

Extended reading notes

Core claim

EgoExoBench is the paper's central discovery: a benchmark that makes cross-view video reasoning measurable. Its eleven subtasks are organized into three dimensions: Ego-Exo Relation (task, action, object, and person matching across views), Ego-Exo View Transition (egocentric wearer identification, direction prediction, and body-part action understanding), and Ego-Exo Temporal Reasoning (action prediction, action ordering, sequence alignment, and skill evaluation). On this benchmark, 13 state-of-the-art MLLMs evaluated in a zero-shot setting score between 29.5 and 48.0 percent on average, with Gemini 2.5 Pro reaching 51.7 percent in the appendix, while deliberate human accuracy on a 330-question subset is 90.1 percent. Models that are strong on single-view benchmarks drop substantially when asked to relate two views. EgoGPT, a model fine-tuned on a large egocentric corpus, gains little over similarly sized general models, and chain-of-thought prompting hurts more than it helps on most subtasks.

Load-bearing premise

The benchmark's validity depends on the assumption that the LLM-based filtering removes every question that can be answered from text alone or from a single view, so that the remaining 7,330 multiple-choice questions genuinely require cross-view visual reasoning.

Editorial extensions

If this is right

  • If the central claim is right, strong single-view video benchmarks do not predict performance on cross-view tasks, so EgoExoBench should be used alongside them when evaluating embodied and assistant models.
  • Egocentric pretraining alone is not a substitute for training objectives that explicitly require associating first- and third-person views.
  • Chain-of-thought-style language reasoning is not the fix; architectures or prompting methods that interleave visual and textual evidence across streams are needed.
  • Providing a reference video from the other viewpoint can improve next-action prediction by up to about 9 points, so paired demonstrations may help future assistants.
  • The benchmark gives the field a concrete 7,330-question yardstick: closing the 42-point gap to deliberate human accuracy is a defined target.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The authors do not measure how often the Qwen2.5-32B filter is wrong; leaking text-solvable questions would inflate model scores and shrink the reported gap, whereas discarding questions that need visual grounding would change what the benchmark measures.
  • A testable extension is to train an MLLM on the benchmark's paired videos with an explicit spatial-alignment objective, such as predicting the egocentric wearer's bounding box in the exocentric frame; the largest gains should appear on Egocentric Wearer Identification and Direction Prediction, the two spatial subtasks with the widest human-model gaps.
  • Because the answer options are deliberately matched for scene and action, a model that learns viewpoint-invariant semantics without true spatial mapping could score well on Relation tasks while still failing View Transition; that dissociation would be a useful diagnostic for future systems.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper introduces EgoExoBench, a benchmark for evaluating multimodal large language models (MLLMs) on video understanding that requires integrating first-person (egocentric) and third-person (exocentric) views. The benchmark contains 7,330 four-choice questions across 11 subtasks grouped into three dimensions: Ego-Exo Relation, Ego-Exo View Transition, and Ego-Exo Temporal Reasoning. Questions are constructed from six public ego-exo datasets using annotation-derived templates, LLM-generated QA from Qwen2.5-32B and DeepSeek-V3, and human annotation. The authors evaluate 13 (or 14, counting the appendix) open- and closed-source MLLMs and report that the best model scores around 48% while deliberate human performance is 90.1%, leaving a 42-point gap. They further find that chain-of-thought prompting does not help and that adding a cross-view reference video helps some tasks but not others. The paper's central claim is that current MLLMs, despite strong single-view performance, lack the ability to semantically align, spatially associate, and temporally reason across ego-exo viewpoints.

Significance. If established, EgoExoBench would fill a genuine gap as the first benchmark targeting cross-view ego-exo reasoning in MLLMs, and the reported human-model gap would be an important result for embodied AI and human-robot collaboration. The paper's strengths include construction from multiple public datasets, a structured task taxonomy, a human baseline, a sanity check that MLLMs can distinguish multiple video inputs, and release of the benchmark. However, the central quantitative claim is currently under-supported: the benchmark lacks matched single-view control conditions, so low scores cannot be confidently attributed to cross-view reasoning rather than to general video difficulty or answer-format artifacts. In addition, the main table contains internal numerical contradictions and omits the best-performing model reported in the appendix. These issues are fixable and the benchmark resource is valuable, but the paper requires substantive revision before its central claim can be accepted.

major comments (4)
  1. [Table 1, Section 4.2, Appendix B.2] The headline human-model gap is not reproducible from the reported numbers. In Table 1, Qwen2.5-VL-72B appears twice, once with Avg 48.5 and once with Avg 44.7; the main text states that Qwen2.5-VL-72B achieves the highest overall accuracy at 47.0%, but Table 1 shows GPT-o4-mini at 48.0 as the best overall model and Qwen2.5-VL-72B at 44.7. Appendix B.2 then reports Gemini 2.5 Pro as best overall at 51.7, yet this model is absent from the main table. Depending on which table is used, the best-model-versus-human gap is approximately 38.4 to 42.1 points, and the claimed best model changes. Please reconcile these numbers, include Gemini 2.5 Pro in the main results, and report a single consistent best-model accuracy.
  2. [Section 3.2 and Section 3.3] The 'Vision-Grounded Filtering' step discards only questions that a text-only LLM can answer from the question text alone (Figure 11/prompt in Appendix A). This does not remove questions that are solvable from a single video view, and for tasks whose options are video clips (TR, AR, OR, PR, EWI, DP, SE, AP) there is no text for the filter to reason over at all. The comparison with single-view benchmarks in Table 5 uses different videos, question formats, and difficulty levels, so it is not a matched control. Without a control in which the same items are presented with only one view (or with same-view distractors), the reported low MLLM scores could reflect general video comprehension difficulty or MCQ artifacts rather than a cross-view reasoning deficit. Please add per-task single-view controls and an analysis of single-view solvability, and use them to substantiate the claim that the benchmark isolates cross-view reasoning.
  3. [Section 3.3.1, Action Relation and Object Relation] The construction of negative candidates for Action Relation is underspecified regarding viewpoint. The text says the ground-truth clip is 'temporally aligned with the query but recorded from a different viewpoint,' while negative candidates are 'selected from different temporal segments of the same video.' If the negatives come from the same viewpoint as the query, then the correct answer is the only candidate from a different viewpoint, and the task reduces to viewpoint identification rather than semantic action matching. Please clarify the viewpoint composition of all options in AR and OR, and verify with an ablation that the tasks cannot be solved by low-level viewpoint or appearance cues alone.
  4. [Section 4.2 and Appendix A.5] The human baseline rests on a small sample: 30 questions per subtask (330 total) answered by two graduate students. The paper reports only a single average (90.1%) with no confidence intervals, no per-annotator breakdown, and no chance-level comparison, even though some subtasks may have uneven answer distributions. This baseline is load-bearing for the central 42-point gap. Please report variance, inter-annotator agreement, and per-subtask human accuracies, and consider increasing the human evaluation set or at least providing error bars.
minor comments (5)
  1. [Throughout] The model name is inconsistent: the main text uses 'GPT-o4-mini' while the reference list and Appendix B.2 use 'GPT-4o mini' (or 'GPT-4o-mini'). Please unify the naming and the count of evaluated models, which is stated as 13 in the abstract but appears to be 14 including Gemini 2.5 Pro.
  2. [Table 5] The EgoExoBench column in Table 5 reports different values for the same models than Table 1 and Table 4 (e.g., Qwen2.5-VL-72B appears as 45.3 versus 44.7; InternVL3-78B as 40.5 versus 41.4). Please ensure that all tables report the same underlying evaluation numbers.
  3. [Section 3.3.2, Egocentric Wearer Identification] The construction text says TF2023 provides 'synchronized egocentric–exocentric image pairs,' but the task is described as video-based. Please clarify whether EWI (and DP) use video clips or static images, and if any subtask is image-based, qualify the 'video understanding' framing accordingly.
  4. [Section 4.3] The chain-of-thought analysis samples 100 questions per subtask and reports per-task accuracy differences without error bars or significance tests. Given the small sample and the magnitude of the observed differences, please add confidence intervals or a significance analysis to support the claim that CoT degrades performance.
  5. [Figure 2] The caption states that correct answers are 'bold-highlighted,' but the figure appears to use green highlighting. Please align the caption with the actual figure formatting.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: the benchmark's answer keys come from annotations and human evaluation, and the central model-deficit claim is not fitted to or defined by the filtering LLMs.

full rationale

The paper's central measurements are model accuracies on MCQs whose correct answers are fixed by public dataset annotations, template construction, human annotation, and human evaluation (Sections 3.2 and A.2-A.5); no answer key is obtained from the evaluated models. The Qwen2.5-32B and DeepSeek-V3 pipeline is used for distractor generation, consistency checking, and text-only filtering, but the reported accuracy numbers are not defined in terms of those LLMs' outputs, so no prediction reduces by construction to a fitted input. Self-citations such as EgoExoLearn [23] as a data source and related-work references like CG-Bench [8] are public, independently available resources and do not carry the paper's conclusion; the paper's empirical comparison is self-contained and uses external single-view benchmarks only as context. The plausible validity concern that the vision-grounded filter removes only text-answerable questions, not single-view-solvable ones, affects construct validity of the cross-view claim but is not circularity: it does not equate any derived result with an input by definition. Reporting inconsistencies (e.g., the duplicated Qwen2.5-VL-72B row in Table 1 and Gemini 2.5 Pro appearing only in the appendix) are correctness and presentation issues, not circular steps.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

The benchmark introduces no fitted constants and no new physical or model entities. Its load-bearing assumptions are about annotation correctness, the effectiveness of LLM-based filtering, the validity of the MCQ format, and the representativeness of the small human baseline.

assumptions (4)
  • domain assumption The six source datasets provide accurate ego-exo pairing and correct annotations such as synchronization, action labels, bounding boxes, and skill scores.
    The QA labels are derived from these annotations throughout Section 3.3. If the underlying pairing or annotations are wrong, the benchmark answers are wrong.
  • domain assumption The Qwen2.5-32B and DeepSeek-V3 based filters remove all text-answerable and ambiguous questions, so the remaining MCQs require visual and cross-view understanding.
    This is the main guarantee that low MLLM scores reflect cross-view reasoning. The filter accuracy is not evaluated or quantified in the paper.
  • domain assumption A four-way multiple-choice format with one correct answer is a valid instrument for measuring cross-view reasoning.
    Section 3.1 adopts MCQ to make evaluation reliable. This assumes the option design does not introduce shortcuts, surface cues, or ordering biases.
  • domain assumption Two graduate students answering a 330-question subset provide a representative human performance baseline.
    Section 4.2 and Appendix A.5 use the 90.1% human figure as the reference point for the human-model gap. The sample is small and the protocol is described inconsistently.

how reviews work

0 comments
Cite this review

Pith. "Pith review of EgoExoBench: A Benchmark for First- and Third-person View Video Understanding in MLLMs." pith.science (2026). https://pith.science/paper/K6L7Y2LH

@misc{pith2026250718342,
  author       = {Pith},
  title        = {Pith review of: EgoExoBench: A Benchmark for First- and Third-person View Video Understanding in MLLMs},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/K6L7Y2LH}},
  note         = {Machine review of arXiv:2507.18342}
}
read the original abstract

Transferring and integrating knowledge across first-person (egocentric) and third-person (exocentric) viewpoints is intrinsic to human intelligence, enabling humans to learn from others and convey insights from their own experiences. Despite rapid progress in multimodal large language models (MLLMs), their ability to perform such cross-view reasoning remains unexplored. To address this, we introduce EgoExoBench, the first benchmark for egocentric-exocentric video understanding and reasoning. Built from publicly available datasets, EgoExoBench comprises over 7,300 question-answer pairs spanning eleven sub-tasks organized into three core challenges: semantic alignment, viewpoint association, and temporal reasoning. We evaluate 13 state-of-the-art MLLMs and find that while these models excel on single-view tasks, they struggle to align semantics across perspectives, accurately associate views, and infer temporal dynamics in the ego-exo context. We hope EgoExoBench can serve as a valuable resource for research on embodied agents and intelligent assistants seeking human-like cross-view intelligence.

Figures

Figures reproduced from arXiv: 2507.18342 by the authors.

Figure 1
Figure 1. The Construction Pipeline of EgoExoBench. [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Example MCQs in EgoExoBench. The correct answers are bold-highlighted. [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Distribution of MCQs in EgoExoBench. Left: number of questions per subtask. Right: [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (24 more)
Figure 4
Figure 4. Figure 4: Relative performance changes after applying CoT. On average, CoT leads to a performance [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]
Figure 5
Figure 5. Figure 5: A failure case of GPT-4o with CoT prompting. The model fails to maintain consistency [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]
Figure 6
Figure 6. Figure 6: Overview of QA distribution in EgoExoBench. [PITH_FULL_IMAGE:figures/full_fig_p015_6.png]
Figure 7
Figure 7. Figure 7: Examples on Ego–Exo Relation. Correct answers are highlighted in green. [PITH_FULL_IMAGE:figures/full_fig_p016_7.png]
Figure 8
Figure 8. Figure 8: Examples on Ego–Exo Relation. Correct answers are highlighted in green. [PITH_FULL_IMAGE:figures/full_fig_p017_8.png]
Figure 9
Figure 9. Figure 9: Examples on Ego–Exo View Transition. Correct answers are highlighted in green. [PITH_FULL_IMAGE:figures/full_fig_p019_9.png]
Figure 10
Figure 10. Figure 10: Prompt for QA generation in the Body Part Action Understanding subtask. [PITH_FULL_IMAGE:figures/full_fig_p020_10.png]
Figure 11
Figure 11. Figure 11: Prompt used to filter out QA pairs that can be correctly answered using text-only input. [PITH_FULL_IMAGE:figures/full_fig_p020_11.png]
Figure 12
Figure 12. Figure 12: Prompt for generating negative options for Action Prediction questions constructed from [PITH_FULL_IMAGE:figures/full_fig_p021_12.png]
Figure 13
Figure 13. Figure 13: Prompt for generating negative options for Action Prediction questions constructed from [PITH_FULL_IMAGE:figures/full_fig_p021_13.png]
Figure 14
Figure 14. Figure 14: Prompt for filtering action pairs in the Action Order subtask to ensure plausible temporal [PITH_FULL_IMAGE:figures/full_fig_p022_14.png]
Figure 15
Figure 15. Figure 15: Prompt for QA generation in the Sequence Alignment subtask. [PITH_FULL_IMAGE:figures/full_fig_p022_15.png]
Figure 16
Figure 16. Figure 16: Examples on Ego–Exo Temporal Reasoning. Correct answers are highlighted in green. [PITH_FULL_IMAGE:figures/full_fig_p023_16.png]
Figure 17
Figure 17. Figure 17: Examples on Ego–Exo Temporal Reasoning. Correct answers are highlighted in green. [PITH_FULL_IMAGE:figures/full_fig_p024_17.png]
Figure 18
Figure 18. Figure 18: Example of the identical video pair identification task proposed in the preliminary [PITH_FULL_IMAGE:figures/full_fig_p026_18.png]
Figure 19
Figure 19. Figure 19: Prompt for extracting the final answer from CoT prediction. [PITH_FULL_IMAGE:figures/full_fig_p027_19.png]
Figure 20
Figure 20. Figure 20: A successful case of GPT-4o’s CoT reasoning on the Task Relation subtask. The model [PITH_FULL_IMAGE:figures/full_fig_p028_20.png]
Figure 21
Figure 21. Figure 21: A successful case of GPT-4o’s CoT reasoning on the Sequence Alignment subtask. The [PITH_FULL_IMAGE:figures/full_fig_p029_21.png]
Figure 22
Figure 22. Figure 22: A failure case of GPT-4o’s CoT reasoning on the Body Part Action Understanding [PITH_FULL_IMAGE:figures/full_fig_p030_22.png]
Figure 23
Figure 23. Figure 23: A failure case of GPT-4o on the Person Relation subtask. The model infers the target [PITH_FULL_IMAGE:figures/full_fig_p031_23.png]
Figure 24
Figure 24. Figure 24: Prompts for the Action Prediction subtask, with and without the reference video. [PITH_FULL_IMAGE:figures/full_fig_p032_24.png]
Figure 25
Figure 25. Figure 25: Prompts for the Skill Evaluation subtask, with and without the reference video. [PITH_FULL_IMAGE:figures/full_fig_p032_25.png]
Figure 26
Figure 26. Figure 26: GPT-4o predictions on the Action Prediction subtask with and without the reference video. [PITH_FULL_IMAGE:figures/full_fig_p033_26.png]
Figure 27
Figure 27. Figure 27: GPT-4o predictions on the Skill Evaluation subtask with and without the reference video. [PITH_FULL_IMAGE:figures/full_fig_p034_27.png]

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 6 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Spatial-IQ: Deconstructing Spatial Intelligence via Hierarchical Capability Tests

    cs.CV 2026-07 conditional novelty 7.0 of 10

    A hierarchical diagnostic benchmark for counting stacked 3D objects shows that MLLMs often reach correct totals without executing the human-like prerequisite sub-tasks, and that training on those sub-tasks markedly im...

  2. Vinci2: Providing Proactive Assistance in Continuous Egocentric Videos

    cs.CV 2026-07 conditional novelty 6.5 of 10

    EgoMemo uses multi-scale temporal summaries, a knowledge graph, and visual archives to decide whether and when to intervene proactively on continuous egocentric video, setting baselines on the new EgoServe benchmark o...

  3. From Synchrony to Sequence: Exo-to-Ego Generation via Interpolation

    cs.CV 2026-04 unverdicted novelty 6.0 of 10

    Interpolating only the video frames between synchronized exo and ego clips already turns discontinuous cross-view generation into continuous sequence modeling and measurably improves diffusion-based Exo2Ego synthesis.

  4. Thinking in Structures: Evaluating Spatial Intelligence in Constraint-Governed Spaces

    cs.CV 2026-02 conditional novelty 6.0 of 10

    A new human-curated benchmark of 1,000 ranking questions on real-world engineering structures shows the best vision-language model reaches 33.6% accuracy where humans reach 91.6%.

  5. The N-Body Problem: Parallel Execution from Single-Person Egocentric Video

    cs.CV 2025-12 conditional novelty 6.0 of 10

    Structured prompting for a VLM, fed with ground-truth spatial zone schedules, can generate more feasible multi-agent parallel executions from single-person egocentric videos.

  6. EgoExo-Con: Exploring View-Invariant Video Temporal Understanding

    cs.CV 2025-10 conditional novelty 6.0 of 10

    Most Video-LLMs answer temporal questions far less consistently when the same event is shown from ego and exo views, and a GRPO variant with a reasoning-similarity reward partially closes the gap.

Reference graph

Works this paper leans on

95 extracted references · 44 canonical work pages · cited by 6 Pith papers

  1. [1]

    Visual-policy learning through multi-camera view to single-camera view knowledge distillation for robot manipulation tasks

    Cihan Acar, Kuluhan Binici, Alp Tekirda˘g, and Yan Wu. Visual-policy learning through multi-camera view to single-camera view knowledge distillation for robot manipulation tasks. IEEE Robotics and Automation Letters, 9(1):691–698, 2023

    2023

  2. [2]

    Claude 3.7 sonnet and claude code, 2025

    Anthropic. Claude 3.7 sonnet and claude code, 2025. URL https://www.anthropic.com/news/ claude-3-7-sonnet

    2025

  3. [3]

    Observational learning

    Albert Bandura. Observational learning. The international encyclopedia of communication, 2008

    2008

  4. [4]

    Siddhant Bansal, Chetan Arora, and C.V . Jawahar. My view is the best view: Procedure learning from egocentric videos. In Eur. Conf. Comput. Vis., 2022

    2022

  5. [5]

    In your place: neuropsychological evidence for altercentric remapping in embodied perspective taking

    Cristina Becchio, Marco Del Giudice, Olga Dal Monte, Luca Latini-Corazzini, and Lorenzo Pia. In your place: neuropsychological evidence for altercentric remapping in embodied perspective taking. Social cognitive and affective neuroscience, 8(2):165–170, 2013

    2013

  6. [6]

    Spatial memory: how egocentric and allocentric combine

    Neil Burgess. Spatial memory: how egocentric and allocentric combine. Trends in cognitive sciences, 10 (12):551–557, 2006

    2006

  7. [7]

    Hourvideo: 1-hour video-language understanding

    Keshigeyan Chandrasegaran, Agrim Gupta, Lea M Hadzic, Taran Kota, Jimming He, Cristóbal Eyzaguirre, Zane Durante, Manling Li, Jiajun Wu, and Fei-Fei Li. Hourvideo: 1-hour video-language understanding. Advances in Neural Information Processing Systems, 37:53168–53197, 2024

    2024

  8. [8]

    Cg-bench: Clue-grounded question answering benchmark for long video understanding

    Guo Chen, Yicheng Liu, Yifei Huang, Yuping He, Baoqi Pei, Jilan Xu, Yali Wang, Tong Lu, and Limin Wang. Cg-bench: Clue-grounded question answering benchmark for long video understanding. arXiv preprint arXiv:2412.12075, 2024

    arXiv 2024

Show all 95 references
  1. [9]

    put myself into your place

    Massimiliano Conson, Elisabetta Mazzarella, Dalila Esposito, Dario Grossi, Nicoletta Marino, Angelo Massagli, and Alessandro Frolli. “put myself into your place”: Embodied simulation and perspective taking in autism spectrum disorders. Autism Research, 8(4):454–466, 2015

    2015

  2. [10]

    Scaling egocentric vision: The epic-kitchens dataset

    Dima Damen, Hazel Doughty, Giovanni Maria Farinella, Sanja Fidler, Antonino Furnari, Evangelos Kazakos, Davide Moltisanti, Jonathan Munro, Toby Perrett, Will Price, et al. Scaling egocentric vision: The epic-kitchens dataset. In Proceedings of the European conference on comput...

    2018

  3. [11]

    Epic-kitchens visor benchmark: Video segmentations and object relations

    Ahmad Darkhalil, Dandan Shan, Bin Zhu, Jian Ma, Amlan Kar, Richard Higgins, Sanja Fidler, David Fouhey, and Dima Damen. Epic-kitchens visor benchmark: Video segmentations and object relations. In Proceedings of the Neural Information Processing Systems (NeurIPS) Track on Datas...

    2022

  4. [12]

    Gemini 2.5 pro, 2025

    DeepMind. Gemini 2.5 pro, 2025. URL https://deepmind.google/models/gemini/pro/

    2025

  5. [13]

    Interacting networks of brain regions underlie human spatial navigation: a review and novel synthesis of the literature

    Arne D Ekstrom, Derek J Huffman, and Michael Starrett. Interacting networks of brain regions underlie human spatial navigation: a review and novel synthesis of the literature. Journal of neurophysiology, 118 (6):3328–3344, 2017

    2017

  6. [14]

    Video-mme: The first-ever comprehensive evaluation benchmark of multi-modal llms in video analysis

    Chaoyou Fu, Yuhan Dai, Yondong Luo, Lei Li, Shuhuai Ren, Renrui Zhang, Zihan Wang, Chenyu Zhou, Yunhang Shen, Mengdan Zhang, et al. Video-mme: The first-ever comprehensive evaluation benchmark of multi-modal llms in video analysis. arXiv preprint arXiv:2405.21075, 2024

    2024 arXiv

  7. [15]

    Complexity-based prompting for multi-step reasoning

    Yao Fu, Hao Peng, Ashish Sabharwal, Peter Clark, and Tushar Khot. Complexity-based prompting for multi-step reasoning. arXiv preprint arXiv:2210.00720, 2022. 10

    2022 arXiv

  8. [16]

    Ego4d: Around the world in 3,000 hours of egocentric video

    Kristen Grauman, Andrew Westbury, Eugene Byrne, Zachary Chavis, Antonino Furnari, Rohit Girdhar, Jackson Hamburger, Hao Jiang, Miao Liu, Xingyu Liu, et al. Ego4d: Around the world in 3,000 hours of egocentric video. In Proceedings of the IEEE/CVF Conference on Computer Vision ...

    2022

  9. [17]

    Ego-exo4d: Understanding skilled human activity from first-and third-person perspectives

    Kristen Grauman, Andrew Westbury, Lorenzo Torresani, Kris Kitani, Jitendra Malik, Triantafyllos Afouras, Kumar Ashutosh, Vijay Baiyya, Siddhant Bansal, Bikram Boote, et al. Ego-exo4d: Understanding skilled human activity from first-and third-person perspectives. In Proceedings...

    2024

  10. [18]

    Ava: A video dataset of spatio-temporally localized atomic visual actions

    Chunhui Gu et al. Ava: A video dataset of spatio-temporally localized atomic visual actions. In IEEE Conf. Comput. Vis. Pattern Recog., pages 6047–6056, 2018. doi: 10.1109/CVPR.2018.00633

    2018

  11. [19]

    Multiple human association and tracking from egocentric and complementary top views

    Ruize Han, Wei Feng, Yujun Zhang, Jiewen Zhao, and Song Wang. Multiple human association and tracking from egocentric and complementary top views. IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(9):5225–5242, 2022. doi: 10.1109/TPAMI.2021.3070562

    2022

  12. [20]

    What is modelled during observational learning? Journal of sports sciences, 25(5):531–545, 2007

    Nicola J Hodges, A Mark Williams, Spencer J Hayes, and Gavin Breslin. What is modelled during observational learning? Journal of sports sciences, 25(5):531–545, 2007

    2007

  13. [21]

    An ego-vision system for discovering human joint attention

    Yifei Huang, Minjie Cai, and Yoichi Sato. An ego-vision system for discovering human joint attention. IEEE Transactions on Human-Machine Systems, 50(4):306–316, 2020

    2020

  14. [22]

    Improving action segmentation via graph-based temporal reasoning

    Yifei Huang, Yusuke Sugano, and Yoichi Sato. Improving action segmentation via graph-based temporal reasoning. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 14024–14034, 2020

    2020

  15. [23]

    Egoexolearn: A dataset for bridging asynchronous ego-and exo-centric view of procedural activities in real world

    Yifei Huang, Guo Chen, Jilan Xu, Mingfang Zhang, Lijin Yang, Baoqi Pei, Hongjie Zhang, Lu Dong, Yali Wang, Limin Wang, et al. Egoexolearn: A dataset for bridging asynchronous ego-and exo-centric view of procedural activities in real world. In Proceedings of the IEEE/CVF Confer...

    2024

  16. [24]

    Vinci: A real-time embodied smart assistant based on egocentric vision-language model

    Yifei Huang, Jilan Xu, Baoqi Pei, Yuping He, Guo Chen, Lijin Yang, Xinyuan Chen, Yaohui Wang, Zheng Nie, Jinyao Liu, et al. Vinci: A real-time embodied smart assistant based on egocentric vision-language model. arXiv preprint arXiv:2412.21080, 2024

    2024 arXiv

  17. [25]

    Gpt-4o system card

    Aaron Hurst, Adam Lerer, Adam P Goucher, Adam Perelman, Aditya Ramesh, Aidan Clark, AJ Ostrow, Akila Welihinda, Alan Hayes, Alec Radford, et al. Gpt-4o system card. arXiv preprint arXiv:2410.21276, 2024

    2024 arXiv

  18. [26]

    Lemma: A multi-view dataset for le arning m ulti-agent m ulti-task a ctivities

    Baoxiong Jia, Yixin Chen, Siyuan Huang, Yixin Zhu, and Song-Chun Zhu. Lemma: A multi-view dataset for le arning m ulti-agent m ulti-task a ctivities. In European Conference on Computer Vision, pages 767–786. Springer, 2020

    2020

  19. [27]

    Egotaskqa: Understanding human tasks in egocentric videos

    Baoxiong Jia, Ting Lei, Song-Chun Zhu, and Siyuan Huang. Egotaskqa: Understanding human tasks in egocentric videos. Advances in Neural Information Processing Systems, 35:3343–3360, 2022

    2022

  20. [28]

    Large language models are zero-shot reasoners

    Takeshi Kojima, Shixiang Shane Gu, Machel Reid, Yutaka Matsuo, and Yusuke Iwasawa. Large language models are zero-shot reasoners. Advances in neural information processing systems, 35:22199–22213, 2022

    2022

  21. [29]

    A neural code for egocentric spatial maps in the human medial temporal lobe

    Lukas Kunz, Armin Brandt, Peter C Reinacher, Bernhard P Staresina, Eric T Reifenstein, Christoph T Weidemann, Nora A Herweg, Ansh Patel, Melina Tsitsiklis, Richard Kempter, et al. A neural code for egocentric spatial maps in the human medial temporal lobe. Neuron, 109(17):2781...

    2021

  22. [30]

    H2o: Two hands manipulating objects for first person interaction recognition

    Taein Kwon, Bugra Tekin, Jan Stühmer, Federica Bogo, and Marc Pollefeys. H2o: Two hands manipulating objects for first person interaction recognition. In Int. Conf. Comput. Vis., 2021

    2021

  23. [31]

    Llava-onevision: Easy visual task transfer

    Bo Li, Yuanhan Zhang, Dong Guo, Renrui Zhang, Feng Li, Hao Zhang, Kaichen Zhang, Peiyuan Zhang, Yanwei Li, Ziwei Liu, et al. Llava-onevision: Easy visual task transfer. arXiv preprint arXiv:2408.03326, 2024

    2024 arXiv

  24. [32]

    igibson 2.0: Object-centric simulation for robot learning of everyday household tasks

    Chengshu Li, Fei Xia, Roberto Martín-Martín, Michael Lingelbach, Sanjana Srivastava, Bokui Shen, Kent Vainio, Cem Gokmen, Gokul Dharan, Tanish Jain, et al. igibson 2.0: Object-centric simulation for robot learning of everyday household tasks. arXiv preprint arXiv:2108.03272, 2021

    2021 arXiv

  25. [33]

    Videochat: Chat-centric video understanding

    KunChang Li, Yinan He, Yi Wang, Yizhuo Li, Wenhai Wang, Ping Luo, Yali Wang, Limin Wang, and Yu Qiao. Videochat: Chat-centric video understanding. arXiv preprint arXiv:2305.06355, 2023. 11

    2023 arXiv

  26. [34]

    Mvbench: A comprehensive multi-modal video understanding benchmark

    Kunchang Li, Yali Wang, Yinan He, Yizhuo Li, Yi Wang, Yi Liu, Zun Wang, Jilan Xu, Guo Chen, Ping Luo, et al. Mvbench: A comprehensive multi-modal video understanding benchmark. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 22195–22...

    2024

  27. [35]

    Ego-exo: Transferring visual representa- tions from third-person to first-person videos

    Yanghao Li, Tushar Nagarajan, Bo Xiong, and Kristen Grauman. Ego-exo: Transferring visual representa- tions from third-person to first-person videos. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 6943–6953, 2021

    2021

  28. [36]

    Deepseek-v3 technical report

    Aixin Liu, Bei Feng, Bing Xue, Bingxuan Wang, Bochao Wu, Chengda Lu, Chenggang Zhao, Chengqi Deng, Chenyu Zhang, Chong Ruan, et al. Deepseek-v3 technical report. arXiv preprint arXiv:2412.19437, 2024

    2024 arXiv

  29. [37]

    Skill transfer learning for autonomous robots and human–robot cooperation: A survey

    Yueyue Liu, Zhijun Li, Huaping Liu, and Zhen Kan. Skill transfer learning for autonomous robots and human–robot cooperation: A survey. Robotics and Autonomous Systems, 128:103515, 2020

    2020

  30. [38]

    Hoi4d: A 4d egocentric dataset for category-level human-object interaction

    Yunze Liu, Yun Liu, Che Jiang, Kangbo Lyu, Weikang Wan, Hao Shen, Boqiang Liang, Zhoujie Fu, He Wang, and Li Yi. Hoi4d: A 4d egocentric dataset for category-level human-object interaction. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pa...

    2022

  31. [39]

    Nvila: Efficient frontier visual language models

    Zhijian Liu, Ligeng Zhu, Baifeng Shi, Zhuoyang Zhang, Yuming Lou, Shang Yang, Haocheng Xi, Shiyi Cao, Yuxian Gu, Dacheng Li, et al. Nvila: Efficient frontier visual language models. arXiv preprint arXiv:2412.04468, 2024

    2024 arXiv

  32. [40]

    Egoschema: A diagnostic benchmark for very long-form video language understanding

    Karttikeya Mangalam, Raiymbek Akshulakov, and Jitendra Malik. Egoschema: A diagnostic benchmark for very long-form video language understanding. Advances in Neural Information Processing Systems, 36:46212–46244, 2023

    2023

  33. [41]

    Howto100m: Learning a text-video embedding by watching hundred million narrated video clips

    Antoine Miech, Dimitri Zhukov, Jean-Baptiste Alayrac, Makarand Tapaswi, Ivan Laptev, and Josef Sivic. Howto100m: Learning a text-video embedding by watching hundred million narrated video clips. In Proceedings of the IEEE/CVF international conference on computer vision, pages ...

    2019

  34. [42]

    Gpt-4o mini: advancing cost-efficient intelligence, 07 2024

    OpenAI. Gpt-4o mini: advancing cost-efficient intelligence, 07 2024. URL https://openai.com/ index/gpt-4o-mini-advancing-cost-efficient-intelligence/

    2024

  35. [43]

    Egovlpv2: Egocentric video-language pre-training with fusion in the backbone

    Shraman Pramanick, Yale Song, Sayan Nag, Kevin Qinghong Lin, Hardik Shah, Mike Zheng Shou, Rama Chellappa, and Pengchuan Zhang. Egovlpv2: Egocentric video-language pre-training with fusion in the backbone. In Proceedings of the IEEE/CVF International Conference on Computer Vis...

    2023

  36. [44]

    Egome: Follow me via egocentric view in real world

    Heqian Qiu, Zhaofeng Shi, Lanxiao Wang, Huiyu Xiong, Xiang Li, and Hongliang Li. Egome: Follow me via egocentric view in real world. arXiv preprint arXiv:2501.19061, 2025

    2025 arXiv

  37. [45]

    Synchronization is all you need: Exocentric-to-egocentric transfer for temporal action segmenta- tion with unlabeled synchronized video pairs, 2023

    Camillo Quattrocchi, Antonino Furnari, Daniele Di Mauro, Mario Valerio Giuffrida, and Giovanni Maria Farinella. Synchronization is all you need: Exocentric-to-egocentric transfer for temporal action segmenta- tion with unlabeled synchronized video pairs, 2023. arXiv:2312.02638

    2023 arXiv

  38. [46]

    The meccano dataset: Understanding human-object interactions from egocentric videos in an industrial-like domain

    Francesco Ragusa, Antonino Furnari, Salvatore Livatino, and Giovanni Maria Farinella. The meccano dataset: Understanding human-object interactions from egocentric videos in an industrial-like domain. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer ...

    2021

  39. [47]

    Home action genome: Cooperative compositional action understanding

    Nishant Rai, Haofeng Chen, Jingwei Ji, Rishi Desai, Kazuki Kozuka, Shun Ishizaka, Ehsan Adeli, and Juan Carlos Niebles. Home action genome: Cooperative compositional action understanding. In IEEE Conf. Comput. Vis. Pattern Recog., 2021

    2021

  40. [48]

    Watch and learn: the cognitive neuroscience of learning from others’ actions

    Richard Ramsey, David M Kaplan, and Emily S Cross. Watch and learn: the cognitive neuroscience of learning from others’ actions. Trends in neurosciences, 44(6):478–491, 2021

    2021

  41. [49]

    Sentence-bert: Sentence embeddings using siamese bert-networks

    Nils Reimers and Iryna Gurevych. Sentence-bert: Sentence embeddings using siamese bert-networks. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing. Association for Computational Linguistics, 11 2019. URL https://arxiv.org/abs/1908.10084

    2019 arXiv

  42. [50]

    A systematic survey of prompt engineering in large language models: Techniques and applications

    Pranab Sahoo, Ayush Kumar Singh, Sriparna Saha, Vinija Jain, Samrat Mondal, and Aman Chadha. A systematic survey of prompt engineering in large language models: Techniques and applications. arXiv preprint arXiv:2402.07927, 2024. 12

    2024 arXiv

  43. [51]

    Sener, D

    F. Sener, D. Chatterjee, D. Shelepov, K. He, D. Singhania, R. Wang, and A. Yao. Assembly101: A large-scale multi-view video dataset for understanding procedural activities. In IEEE Conf. Comput. Vis. Pattern Recog., 2022

    2022

  44. [52]

    Self-supervised disentangled representation learning for third-person imitation learning

    Jinghuan Shang and Michael S Ryoo. Self-supervised disentangled representation learning for third-person imitation learning. In 2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 214–221. IEEE, 2021

    2021

  45. [53]

    Third-person visual imitation learning via decoupled hierarchical controller

    Pratyusha Sharma, Deepak Pathak, and Abhinav Gupta. Third-person visual imitation learning via decoupled hierarchical controller. Advances in Neural Information Processing Systems, 32, 2019

    2019

  46. [54]

    Actor and observer: Joint modeling of first and third-person videos

    Gunnar A Sigurdsson, Abhinav Gupta, Cordelia Schmid, Ali Farhadi, and Karteek Alahari. Actor and observer: Joint modeling of first and third-person videos. In proceedings of the IEEE conference on computer vision and pattern recognition, pages 7396–7404, 2018

    2018

  47. [55]

    Ego4d goal-step: Toward hierarchical understanding of procedural activities

    Yale Song, Eugene Byrne, Tushar Nagarajan, Huiyu Wang, Miguel Martin, and Lorenzo Torresani. Ego4d goal-step: Toward hierarchical understanding of procedural activities. Adv. Neural Inform. Process. Syst., 2024

    2024

  48. [56]

    Ucf101: A dataset of 101 human actions classes from videos in the wild, 2012

    Khurram Soomro, Amir Roshan Zamir, and Mubarak Shah. Ucf101: A dataset of 101 human actions classes from videos in the wild, 2012. arXiv:1212.0402

    2012 arXiv

  49. [57]

    Qwen2.5-vl, January 2025

    Qwen Team. Qwen2.5-vl, January 2025. URL https://qwenlm.github.io/blog/qwen2.5-vl/

    2025

  50. [58]

    Learning from semantic alignment between unpaired multiviews for egocentric video recognition

    Qitong Wang, Long Zhao, Liangzhe Yuan, Ting Liu, and Xi Peng. Learning from semantic alignment between unpaired multiviews for egocentric video recognition. In Int. Conf. Comput. Vis., 2023

    2023

  51. [59]

    Lvbench: An extreme long video understanding benchmark

    Weihan Wang, Zehai He, Wenyi Hong, Yean Cheng, Xiaohan Zhang, Ji Qi, Xiaotao Gu, Shiyu Huang, Bin Xu, Yuxiao Dong, et al. Lvbench: An extreme long video understanding benchmark. arXiv preprint arXiv:2406.08035, 2024

    2024 arXiv

  52. [60]

    Self-consistency improves chain of thought reasoning in language models

    Xuezhi Wang, Jason Wei, Dale Schuurmans, Quoc Le, Ed Chi, Sharan Narang, Aakanksha Chowdhery, and Denny Zhou. Self-consistency improves chain of thought reasoning in language models. arXiv preprint arXiv:2203.11171, 2022

    2022 arXiv

  53. [61]

    See what i see: Enabling user-centric robotic assistance using first-person demonstrations

    Yeping Wang, Gopika Ajaykumar, and Chien-Ming Huang. See what i see: Enabling user-centric robotic assistance using first-person demonstrations. In Proceedings of the 2020 ACM/IEEE International Conference on Human-Robot Interaction, pages 639–648, 2020

    2020

  54. [62]

    Chain-of-thought prompting elicits reasoning in large language models

    Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Fei Xia, Ed Chi, Quoc V Le, Denny Zhou, et al. Chain-of-thought prompting elicits reasoning in large language models. Advances in neural information processing systems, 35:24824–24837, 2022

    2022

  55. [63]

    Incomplete multi-view domain adaptation via channel enhancement and knowledge transfer

    Haifeng Xia, Pu Wang, and Zhengming Ding. Incomplete multi-view domain adaptation via channel enhancement and knowledge transfer. In Eur. Conf. Comput. Vis., 2022

    2022

  56. [64]

    Next-qa: Next phase of question-answering to explaining temporal actions

    Junbin Xiao, Xindi Shang, Angela Yao, and Tat-Seng Chua. Next-qa: Next phase of question-answering to explaining temporal actions. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 9777–9786, 2021

    2021

  57. [65]

    Can i trust your answer? visually grounded video question answering

    Junbin Xiao, Angela Yao, Yicong Li, and Tat-Seng Chua. Can i trust your answer? visually grounded video question answering. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 13204–13214, 2024

    2024

  58. [66]

    Pov: Prompt-oriented view-agnostic learning for egocentric hand-object interaction in the multi-view world

    Boshen Xu, Sipeng Zheng, and Qin Jin. Pov: Prompt-oriented view-agnostic learning for egocentric hand-object interaction in the multi-view world. In ACM Int. Conf. Multimedia, pages 2807–2816, 2023

    2023

  59. [67]

    Retrieval-augmented egocentric video captioning

    Jilan Xu, Yifei Huang, Junlin Hou, Guo Chen, Yuejie Zhang, Rui Feng, and Weidi Xie. Retrieval-augmented egocentric video captioning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 13525–13536, 2024

    2024

  60. [68]

    Egoexo-gen: Ego-centric video prediction by watching exo-centric videos

    Jilan Xu, Yifei Huang, Baoqi Pei, Junlin Hou, Qingqiu Li, Guo Chen, Yuejie Zhang, Rui Feng, and Weidi Xie. Egoexo-gen: Ego-centric video prediction by watching exo-centric videos. arXiv preprint arXiv:2504.11732, 2025

    2025 arXiv

  61. [69]

    Advancing high-resolution video-language representation with large-scale video transcriptions

    Hongwei Xue, Tiankai Hang, Yanhong Zeng, Yuchong Sun, Bei Liu, Huan Yang, Jianlong Fu, and Baining Guo. Advancing high-resolution video-language representation with large-scale video transcriptions. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recog...

    2022

  62. [70]

    An Yang, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chengyuan Li, Dayiheng Liu, Fei Huang, Haoran Wei, et al. Qwen2. 5 technical report. arXiv preprint arXiv:2412.15115, 2024

    2024 arXiv

  63. [71]

    Thinking in space: How multimodal large language models see, remember, and recall spaces.arXiv preprint arXiv:2412.14171, 2024

    Jihan Yang, Shusheng Yang, Anjali W Gupta, Rilyn Han, Li Fei-Fei, and Saining Xie. Thinking in space: How multimodal large language models see, remember, and recall spaces.arXiv preprint arXiv:2412.14171, 2024

    2024 arXiv

  64. [72]

    Egolife: Towards egocentric life assistant

    Jingkang Yang, Shuai Liu, Hongming Guo, Yuhao Dong, Xiamengwei Zhang, Sicheng Zhang, Pengyun Wang, Zitang Zhou, Binzhu Xie, Ziyue Wang, et al. Egolife: Towards egocentric life assistant. arXiv preprint arXiv:2503.03803, 2025

    2025

  65. [73]

    Interact before align: Leveraging cross-modal knowledge for domain adaptive action recognition

    Lijin Yang, Yifei Huang, Yusuke Sugano, and Yoichi Sato. Interact before align: Leveraging cross-modal knowledge for domain adaptive action recognition. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2022

    2022

  66. [74]

    Fine-grained affordance annotation for egocentric hand-object interaction videos

    Zecheng Yu, Yifei Huang, Ryosuke Furuta, Takuma Yagi, Yusuke Goutsu, and Yoichi Sato. Fine-grained affordance annotation for egocentric hand-object interaction videos. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 2023

    2023

  67. [75]

    Video instruction tuning with synthetic data

    Yuanhan Zhang, Jinming Wu, Wei Li, Bo Li, Zejun Ma, Ziwei Liu, and Chunyuan Li. Video instruction tuning with synthetic data. arXiv preprint arXiv:2410.02713, 2024

    2024 arXiv

  68. [76]

    Fusing personal and environmental cues for identification and segmentation of first-person camera wearers in third-person views

    Ziwei Zhao, Yuchen Wang, and Chuhua Wang. Fusing personal and environmental cues for identification and segmentation of first-person camera wearers in third-person views. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 16477–16487, 2024

    2024

  69. [77]

    Progressive-hint prompting improves reasoning in large language models

    Chuanyang Zheng, Zhengying Liu, Enze Xie, Zhenguo Li, and Yu Li. Progressive-hint prompting improves reasoning in large language models. arXiv preprint arXiv:2304.09797, 2023

    2023 arXiv

  70. [78]

    Mlvu: A comprehensive benchmark for multi-task long video understanding

    Junjie Zhou, Yan Shu, Bo Zhao, Boya Wu, Shitao Xiao, Xi Yang, Yongping Xiong, Bo Zhang, Tiejun Huang, and Zheng Liu. Mlvu: A comprehensive benchmark for multi-task long video understanding. arXiv preprint arXiv:2406.04264, 2024

    2024 arXiv

  71. [79]

    Internvl3: Exploring advanced training and test-time recipes for open-source multimodal models

    Jinguo Zhu, Weiyun Wang, Zhe Chen, Zhaoyang Liu, Shenglong Ye, Lixin Gu, Yuchen Duan, Hao Tian, Weijie Su, Jie Shao, et al. Internvl3: Exploring advanced training and test-time recipes for open-source multimodal models. arXiv preprint arXiv:2504.10479, 2025. A Additional detai...

    2025 arXiv

  72. [80]

    Each QA pair should include one question, four answer options, and one correct answer

  73. [81]

    The question must clearly address details from both descriptions, particularly focusing on the roles of the body parts, without directly referencing specific actions or objects to prevent giving away hints

  74. [82]

    Ensure there is one correct answer, and the other three options should be plausible but incorrect to increase the difficulty

  75. [83]

    Use 'Video 1' to refer Video 1, use 'Video 2' to refer Video 2

  76. [84]

    Question

    The output must be in JSON format as follows: ```json { "Question": "The question text goes here", "Options": { "A": "Option A text", "B": "Option B text", "C": "Option C text", "D": "Option D text" }, "Answer": "Correct option letter (e.g., 'A')" } ``` Example Input: Video 1:...

  77. [85]

    Providing an incorrect sequence of actions

  78. [86]

    Video 1" and the second video as

    Making false statements about the presence or absence of specific actions; Accuracy: Avoid vague or overly subjective questions. Do not introduce actions or details that are not present in the provided sequences. Naming Convention: Refer to the first video as "Video 1" and the...

  79. [87]

    The person appears to be preparing materials on the table

  80. [88]

    They seem to be handling a test tube, possibly doing a covering or sealing action

  81. [89]

    There's a focus on opening or handling some instruction or paper

  82. [90]

    Video 2:

    The testing plate is being unwrapped and possibly manipulated. Video 2:

  83. [91]

    The person is seen handling materials on a table

  84. [92]

    They unwrap the testing plate and handle the test tube

  85. [93]

    Conclusion: After reviewing the sequences, Option B correctly describes the actions

    After unwrapping, they perform actions with the test tube, such as covering it. Conclusion: After reviewing the sequences, Option B correctly describes the actions. B. In Video 1, the test tube is covered before unwrapping the testing plate, whereas in Video 2, the test tube i...

  86. [94]

    Query Video Analysis: The query video features a person annotated with a bounding box who is dressed in blue and is standing in a more distant position relative to others

  87. [95]

    Video 2: The person within the bounding box is positioned at the bottom of the scene, facing upwards, and is in a different area relative to others

    Candidate Videos Analysis: Video 1: The person within the bounding box is located in the middle area, positioned closer to others, and is facing towards the lower part of the image. Video 2: The person within the bounding box is positioned at the bottom of the scene, facing up...

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.