Pith. sign in

REVIEW 4 major objections 5 minor 53 references

HAS: Highlight-guided Attention Steering for Multimodal LLM Video Summarization

T0 review · 4 major / 5 minor · reviewed 2026-08-01 · deepseek-v4-flash

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

Pith's one-line read A frozen video LLM can summarize better when guided by a continuous highlight curve instead of a discrete keyframe selection.

desk verdict A lightweight steering idea worth a serious look, but the temporal-salience table as written appears to credit the highlight prior rather than the steering itself. read the letter →

arxiv 2607.17994 v1 pith:3XR27WGM submitted 2026-07-20 cs.CV cs.AI

Rui Chu , Yingjie Lao This is my paper
classification cs.CVcs.AI
keywords videosummarizationmultimodalLLMattentionsteeringinference-timecontrolhighlightdetectiontemporalsaliencefactualgroundinglong-form
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

This paper tries to establish that video summarization by a multimodal LLM improves when the model is steered at inference time by a continuous frame-level highlight score, rather than by first selecting a few keyframes and summarizing only those. The authors argue that hard selection discards context and breaks coherence, while soft guidance keeps the full video in front of the model and simply biases its cross-attention toward relevant moments. If correct, open-source video MLLMs can be made more faithful and coherent summarizers without any weight updates, just by adding a small attention bias derived from an existing highlight detector. The paper reports gains in temporal importance ranking, cross-modal consistency, and factual grounding on long-form scientific talks.

What carries the argument

The central mechanism is the steering vector V, built from a query-conditioned highlight distribution h = [h_1, ..., h_T] produced by an off-the-shelf highlight generator, then temporally interpolated, min-max normalized to [0, 1], and lifted to token level by repeating each frame score over the P visual tokens of that frame. The log transform makes V usable as an additive bias on attention logits. During decoding, for each selected cross-attention head, the pre-softmax logits are modified row-wise as A_i,: + g_lm β_lm V, where g_lm = σ(a_lm) is a learned gate and β_lm a head-wise strength; only the gates and strengths are optimized on a validation set, while the MLLM stays frozen. This conv

What would settle it

Run HAS with three steering vectors on the same videos: the real highlight prior, a uniform vector, and a scrambled vector. If uniform or scrambled vectors match the real-prior version on temporal salience and factual-consistency metrics, the reported gains come from the injection machinery rather than the highlight content; if an oracle human-annotation prior clearly outperforms the off-the-shelf prior, the unmeasured quality of H is the decisive factor.

Watch

Extended reading notes

Core claim

HAS claims that a frozen video MLLM M, guided by a steering vector V built from a calibrated highlight distribution h = H(F, q), produces better summaries O* = M_V(F, q) than score-then-select pipelines. The steering vector is constructed by lifting each frame's normalized highlight score to all P visual tokens extracted from that frame, V = log(Repeat(h, P) + ε), and adding it as a gated additive bias to the pre-softmax attention logits of a selected set of cross-attention heads during every decoding step. The intervention is soft: low-scored frames receive less attention, not zero, so the model retains dispersed evidence while concentrating on highlights. The authors present the method as

Load-bearing premise

The central claim rests on the reliability of the off-the-shelf highlight generator H: if its frame scores do not faithfully reflect what a summary should attend to, then adding them to a frozen model's attention logits injects the generator's noise into every decoding step.

Editorial extensions

If this is right

  • Soft steering retains low-scored frames as weak attention targets, so summaries can keep using dispersed evidence that hard selection would irreversibly remove under a fixed frame budget.
  • Because the backbone is frozen and only a small gate/strength policy is learned, the method can be layered on any compatible open-source video MLLM without retraining.
  • The reported zero-shot transfer from one summarization benchmark to another suggests the steering mechanism generalizes across video distributions without target-domain adaptation.
  • Improvements in factual consistency and video–text alignment on long-form talks imply that attention-level control can reduce unsupported details in generated summaries.

Reading between the lines

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

  • An implication the paper leaves implicit: the same steering recipe could be reused for any future video MLLM, so the long-term ceiling of the approach is set by the quality of the highlight prior, not by the steering injection itself.
  • A testable extension is to replace the single global highlight curve with per-section or per-query curves, steering different parts of the summary toward different evidence; the paper's own future-work section hints at a coarse-to-fine policy.
  • The method converts an explainability/saliency signal into a control signal, a pattern that could transfer to other long-context modalities where a global relevance curve is available, such as audio or document summarization.
  • Because the gating policy is tuned on reference summaries, the method may inherit the style of the validation set; zero-shot deployment to domains with very different summary conventions might require recalibrating the gate parameters.
Share X Bluesky LinkedIn Reddit HN

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 proposes HAS, an inference-time attention-steering method for multimodal LLM video summarization. Given a video and a query, an off-the-shelf highlight generator produces a raw temporal salience curve; HAS calibrates this curve into a smooth frame-level highlight distribution, lifts it into a token-level log-space steering vector V = log(Repeat(h,P)+ε), and injects a gated additive bias into selected cross-attention heads of a frozen video MLLM. The head-wise gates and steering strengths are chosen by minimizing teacher-forcing summary NLL on a held-out validation set, with a sparsity regularizer. The paper reports gains on SumMe/TVSum, VideoXum, VISTA, and MR.HiSum, and claims that HAS improves temporal salience agreement, cross-modal consistency, and grounded long-form summarization without updating the backbone.

Significance. The proposed mechanism is clean and non-circular: the highlight prior comes from an external generator, and the steering policy in Eq. (4) is optimized on validation summary NLL rather than on the benchmark metrics. If the claims hold, the contribution is a useful plug-and-play, backbone-agnostic inference-time module with broad benchmark coverage across several open-source video MLLMs. The derivation of Eqs. (2)–(4) is explicit and the method is falsifiable in principle. However, the evidence for the temporal-salience claim is currently confounded by the external prior, and the external highlight generator is never named or evaluated. These issues are central to the paper's main claims and need to be addressed before the results can be accepted.

major comments (4)
  1. [§4.2, Table 1] The caption reports rank correlation between "predicted importance trajectories" and human annotations, but the manuscript never defines what HAS's predicted importance trajectory is. Algorithm 1 returns only V, a deterministic log/Repeat transform of the calibrated highlight prior h; Fig. 4(B) further states that post-steering attention mass "closely follows v." Thus, if the curve scored is V or the attention mass, the reported τ/ρ largely measures the external generator H, not the steering operation in Eq. (3). If it is something else, it is not defined. No ablation (e.g., random, uniform, or shuffled V, or H alone) isolates the steering contribution, so the §4.2 claim that "inference-time steering … improves global salience ordering" is unsupported. Please specify the evaluated trajectory, report H-only, and add prior-quality ablations.
  2. [§3.3, Algorithm 1] H(F,q) is load-bearing but never named, specified, or evaluated on the evaluation datasets. Since V is built directly from H, the injected bias in Eq. (3) adds a scaled version of H's output to attention logits at every decoding step, and §7 concedes that steering quality depends on the reliability of the initial highlight estimate. The paper should name the generator, report its standalone accuracy, and explicitly measure how HAS's metrics vary when H is replaced by corrupted or random priors. Without this, it is impossible to attribute observed gains to the attention-steering policy rather than to the external prior.
  3. [§3.5, Eq. (4) and §4.1] The steering policy {a,β} is fit by minimizing teacher-forcing NLL of reference summaries on a "held-out calibration/validation set," but the experimental section does not state the split, the dataset, or whether one policy is reused across benchmark families. This matters because the method is advertised as an inference-time plug-in; per-dataset validation-set tuning would weaken the zero-shot/generalization claims. Table 4 also says models are "trained on SumMe" even though HAS is a training-free intervention; the exact transfer protocol for HAS must be clarified.
  4. [§4.3, Fig. 2] The "hard selection baseline" is not identified, the budget T schedule is not defined, and the computation of fact-unit recall is not described. The central coverage claim—that soft steering defers saturation relative to discrete extraction—cannot be checked without this protocol. Please specify the baseline(s), budget grid, and metric definitions.
minor comments (5)
  1. [Abstract / Fig. 1 caption] There are several language issues: "lost important information" should be "lose important information," "continual attention guidance" should be "continuous attention guidance," and "peace time steps" should be "non-highlight time steps."
  2. [Table 4] τ and ρ are reported as identical to two decimals for every method (e.g., 0.440/0.440 for LLMVS and 0.45/0.45 for HAS). This is suspicious for rank-correlation measures on the same data; please verify the underlying rankings and report standard deviations or significance tests.
  3. [Table 3] "HAS (ours, averaged on selected backbones)" is vague. Specify which backbones are included in the average and why those are selected; otherwise the comparison to individual zero-shot and fine-tuned baselines is hard to interpret.
  4. [Fig. 3] Only an "averaged" ΔFactVC is shown without per-backbone numerical values. Include a small table with the individual gains for each backbone and the number of runs used for averaging.
  5. [§1, reference [48]] The sentence "Recent multimodal large language models (M-LLMs) [48]" cites a deep-learning video summarization review, not a reference on MLLMs. The citation appears mismatched and should be corrected.

Circularity Check

1 steps flagged · score 6.0 of 10

Table 1's temporal-salience 'prediction' is the calibrated highlight prior (or attention that 'closely follows' it), so that headline gain is inherited from the external H rather than from the steering mechanism.

  1. self definitional [Table 1 caption (§4.1); Fig. 4(B) (§4.5); Algorithm 1 (§3.3); Eqs. (2)–(3) (§3.4–3.5)]
    "Rank correlation (τ /ρ; higher is better) between predicted importance trajectories and human annotations on V2V summarization. ... Calibration converts the noisy raw highlight prior into a temporally coherent distribution v, and the post-steering attention mass closely follows v, indicating effective injection of the prompt-conditioned control signal into inference-time attention."

    Algorithm 1's only trajectory is h = MinMaxNorm(ĥ) with ĥ = H(F, q); Eq. (2) sets V = log(Repeat(h, P) + ε) and Eq. (3) adds V to attention logits. Fig. 4(B) states that post-steering attention mass 'closely follows v', where v is the calibrated highlight prior h. Table 1 scores 'predicted importance trajectories' against human annotations. Thus the reported τ/ρ are inherited from the off-the-shelf highlight generator H, i.e., the method's own input is being evaluated as the method's prediction. No ablation with a random, shuffled, or uniform prior, and no standalone evaluation of H, isolates any contribution from the attention-steering mechanism to temporal salience.

full rationale

The paper's core summarization pipeline is not circular in its derivation: the steering policy {a, β} is optimized on validation summary NLL (Eq. 4), not on the benchmark metrics, and the highlight prior comes from an external, unnamed off-the-shelf generator H. The self-citations present in the reference list are to related robustness/security work and are not load-bearing for the method. However, one of the three headline claims—'improving temporal salience agreement'—rests entirely on Table 1, whose 'predicted importance trajectories' are never defined. Internal evidence (Fig. 4B, Eq. 2, Algorithm 1) shows that the trajectory used for this evaluation is either the calibrated highlight prior h or an attention mass that 'closely follows' h. In either case, the temporal-salience result reduces by construction to scoring the input prior H, not to the proposed attention-steering mechanism. This is a partial circularity: the summarization-quality claims in Tables 2 and 3 are genuine output evaluations, but the temporal-salience claim is an input-as-output evaluation.

Assumptions & free parameters 7 free parameters · 4 assumptions · 0 invented entities

The paper contributes smoothing/normalization of an external highlight curve (Algorithm 1), a log-space bias construction (Eqs. 2-3), and a validation-set-tuned head-selection policy (Eq. 4). All semantic content of the steering signal comes from the unnamed off-the-shelf generator H; the backbone is frozen. Seven free hyperparameters controlling the injection are unreported. No invented entities are introduced: the steering vector V = log(Repeat(h,P)+epsilon) is a deterministic construction, not a postulated entity.

free parameters (7)
  • Per-head gate logits a_{l,m}
    Sigmoid gates in Eq. (3); optimized on validation-set summary NLL via Eq. (4) with Adam; values never reported.
  • Per-head steering strengths beta_{l,m}
    Additive-bias magnitudes in Eq. (3); optimized with the gates; values never reported.
  • Sparsity weight lambda_s
    Regularizer in Eq. (4) trading summary likelihood against the number of activated heads; value never reported.
  • Stability constant epsilon
    Log-space floor in Eq. (2) to avoid collapse to -infinity; value never reported.
  • Candidate head set S
    Which (layer, head) pairs receive the bias is never specified, though it determines the locality of the intervention.
  • Temporal calibration kernel
    Smoothing width used in the Section 3.3 calibration step is never specified.
  • Tokens per frame P
    Eq. (2) repeats each frame score over P visual tokens; P depends on the backbone's tubelet encoding and is not stated.
assumptions (4)
  • domain assumption Off-the-shelf highlight scores are a valid, query-conditioned summary-relevance prior
    Algorithm 1 and Section 3.3 take h_hat = H(F, q) as the semantic source of the steering signal; Section 7 concedes that steering quality depends on the initial highlight estimate.
  • domain assumption Additive log-space bias on selected cross-attention logits steers evidence usage without breaking generation
    Eq. (3) modifies pre-softmax logits; effectiveness is borrowed from cited attention-biasing/PASTA/FarSight works (Section 2.3) plus the Fig. 4(B) visualization.
  • standard math Teacher-forcing NLL of reference summaries is the right objective for steering-policy fitting
    Eq. (4) minimizes L_NLL plus a sparsity term; standard practice, but it means the policy is fit to the target task's summaries (see the training-free framing red flag).
  • domain assumption Steered attention mass correlates with factual grounding (FactVC)
    Implied by Fig. 4(B) and the FactVC gains in Table 3; no direct causal measurement is provided.

how reviews work

0 comments
Cite this review

Pith. "Pith review of HAS: Highlight-guided Attention Steering for Multimodal LLM Video Summarization." pith.science (2026). https://pith.science/paper/3XR27WGM

@misc{pith2026260717994,
  author       = {Pith},
  title        = {Pith review of: HAS: Highlight-guided Attention Steering for Multimodal LLM Video Summarization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3XR27WGM}},
  note         = {Machine review of arXiv:2607.17994}
}
read the original abstract

Video understanding has become more and more important with the growth of Artificial Intelligence (AI) for video generation. Recently, Multimodal Large Language Model(M-LLM) has shown its capability in video understanding. Video summarization, a specific domain of video understanding, has proven its importance for efficient navigation and retrieval. Both video understanding and video summarization require a good selection of key frames in a video. Current video summarization methods heavily focus on the selected key frames and correlated segment captions. However, existing approaches overlook the perspective of treating the importance of the frames globally. We argue that using discrete selected frames for summarization will not only reduce the understanding coherence, but also lost important information in the video, as well as wasting the original capacity of the MLLMs. In this paper, we propose HAS, a Highlight-guided Attention Steering method for video summarization. We consider a challenging but practical setting where the video given to MLLMs for summarize should be continuous but with highlight guidance. HAS mainly consists of two parts: The first part is to find a continuous frame-level highlight distribution for the video globally. The second part is to apply the highlight distribution as an attention steering vector for the MLLM, targeting a better understanding of the video, and thus during the model inference time, putting more attention on the highlighted frames, while avoiding lost entire information on less highlighted frames through putting less attention instead of forgetting them. We evaluated HAS on a variety of benchmarks, and it has shown convincing performance in video summarization.

Figures

Figures reproduced from arXiv: 2607.17994 by the authors.

Figure 1
Figure 1. Top: Unlike prior works assign discrete importance scores and hard-select a few highlights before summarizing, HASpreserves the full video context and treats highlighting as continual attention guidance through steering vector. Bottom: While smoothly bias a frozen video MLLM toward highlight moments, HAS does not neglect the peace time steps, better exploiting model capacity while reducing missed evidence for more c… view at source ↗
Figure 2
Figure 2. Coverage vs. budget on VISTA. (a) ROUGE-Lsum and (b) transcript fact-unit recall versus T. Hard selection saturates early, while HASkeeps improving at mid-to-high budgets, indicating reduced information loss under fixed-budget inference. trend upward without sacrificing overall summary quality. This suggests that calibrated temporal steering encourages evidence usage during generation, re￾ducing unsupported details … view at source ↗
Figure 3
Figure 3. Gains across open-source MLLMs. The improvement (Averaged) in groundedness on VISTA after adding HAS to all selected open-source video MLLM backbone. Positive values indicate that HAS increases factual consistency. the same zero-shot protocol. This directly addresses the single-backbone concern: the gains are not tied to a particular architecture or training recipe, but stem from the inference-time steering mechanis… view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Qualitative and mechanistic visualizations of HAS 5 Conclusion In this work, we presented HAS, a highlight-guided attention steering frame￾work for multimodal LLM video summarization. Unlike prior pipelines that rely on discrete highlight selection, HAS turns query-con…

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

53 extracted references · 2 linked inside Pith

  1. [1]

    In: Proceedings of the IEEE International Symposium on Multimedia (ISM) (2021)

    Apostolidis, E., Balaouras, G., Mezaris, V., Patras, I.: Combining global and local attention with positional encoding for video summarization. In: Proceedings of the IEEE International Symposium on Multimedia (ISM) (2021)

    2021

  2. [2]

    Argaw, D.M., Yoon, S., Caba Heilbron, F., Deilamsalehy, H., Bui, T., Wang, Z., Dernoncourt, F., Chung, J.S.: Scaling up video summarization pretraining with largelanguagemodels.In:IEEE/CVFConferenceonComputerVisionandPattern Recognition (CVPR) (2024),

    2024

  3. [3]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

    Bae, K., Kim, J., Lee, S., Lee, S., Lee, G., Choi, J.: Mash-vlm: Mitigating action- scene hallucination in video-llms through disentangled spatial-temporal represen- tations. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). pp. 13744–13753 (2025)

    2025

  4. [4]

    In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR)

    Bargal, S.A., Zunino, A., Kim, D., Zhang, J., Murino, V., Sclaroff, S.: Excitation backprop for rnns. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). pp. 1440–1449 (2018)

    2018

  5. [5]

    In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

    Buch, S., Eyzaguirre, C., Gaidon, A., Wu, J., Fei-Fei, L., Niebles, J.C.: Revisiting the "video" in video-language understanding. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 2917–2927 (2022)

    2022

  6. [6]

    Chen, G., Huang, Y., Xu, J., Pei, B., Wang, J., Chen, Z., Li, Z., Lu, T., Wang, L.: Video mamba suite: State space model as a versatile alternative for video under- standing. Int. J. Comput. Vis.134(1), 20 (2026). ,

    2026

  7. [7]

    Advances in Neural In- formation Processing Systems38, 14152–14183 (2025)

    Chu, R., Zhao, B., Jiang, H., Aeron, S., Lao, Y.: Bam-icl: Causal hijacking in- context learning with budgeted adversarial manipulation. Advances in Neural In- formation Processing Systems38, 14152–14183 (2025)

    2025

  8. [8]

    Chu, R., Zhao, B., Le, T.Q.H., Hoang, D.C., Lin, H., Li, P., Zhao, W., Doan, K.D., Lao, Y.: Debiasrag: A tuning-free path to fair generation in large language models through retrieval-augmented generation (2026),

    2026

Show all 53 references
  1. [9]

    In: Computer Vision – ACCV 2018 Workshops

    Fajtl, J., Sadeghi Sokeh, H., Argyriou, V., Monekosso, D., Remagnino, P.: Sum- marizing videos with attention. In: Computer Vision – ACCV 2018 Workshops. Lecture Notes in Computer Science, vol. 11367, pp. 39–54. Springer (2019)

    2018

  2. [10]

    arXiv preprint arXiv:2503.16032 (2025)

    Fan,S.,Guo,M.H.,Yang,S.:Agentickeyframesearchforvideoquestionanswering. arXiv preprint arXiv:2503.16032 (2025)

    2025 arXiv

  3. [11]

    CoRRabs/2506.02733(2025)

    Feng, X., Zou, K., Cen, C., Huang, T., Guo, H., Huang, Z., Zhao, Y., Zhang, M., Zheng, Z., Wang, D., Zou, Y., Li, D.: Linkto-anime: A 2d animation optical flow dataset from 3d model rendering. CoRRabs/2506.02733(2025). ,

    2025 arXiv

  4. [12]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)

    Fong, R., Patrick, M., Vedaldi, A.: Understanding deep networks via extremal perturbations and smooth masks. In: Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV). pp. 2950–2958 (2019)

    2019

  5. [13]

    In: Proceedings of the 29th ACM/IEEE International Symposium on Low Power Electronics and Design

    Fu, Z., Avaliani, A., Donato, M.: Heterogeneous memory integration and optimiza- tion for energy-efficient multi-task nlp edge inference. In: Proceedings of the 29th ACM/IEEE International Symposium on Low Power Electronics and Design. p. 1–6. ISLPED ’24, Association for Compu...

    2024

  6. [14]

    In: Proceedings of the IEEE International Conference on Multimedia and Expo (ICME) (2021)

    Ghauri, J.A., Hakimov, S., Ewerth, R.: Supervised video summarization via multi- ple feature sets with parallel attention. In: Proceedings of the IEEE International Conference on Multimedia and Expo (ICME) (2021)

    2021

  7. [15]

    In: Computer Vision – ECCV 2014

    Gygli, M., Grabner, H., Riemenschneider, H., Van Gool, L.: Creating summaries from user videos. In: Computer Vision – ECCV 2014. Lecture Notes in Computer Science, vol. 8695, pp. 505–520. Springer (2014) HAS: Attention Steering for Video Summarization 17

    2014

  8. [16]

    In: IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2025, Nashville, TN, USA, June 11-15, 2025

    Han, Y., Zhao, B., Chu, R., Luo, F., Sikdar, B., Lao, Y.: Uibdiffusion: Universal imperceptible backdoor attack for diffusion models. In: IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2025, Nashville, TN, USA, June 11-15, 2025. pp. 19186–19196. Computer ...

    2025

  9. [17]

    Proceedings of the AAAI Conference on Artificial Intelligence 36(10), 10738–10748 (2022)

    Hazarika, D., Namazifar, M., Hakkani-Tür, D.: Attention biasing and context aug- mentation for zero-shot control of encoder-decoder transformers for natural lan- guage generation. Proceedings of the AAAI Conference on Artificial Intelligence 36(10), 10738–10748 (2022)

    2022

  10. [18]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2023)

    He, B., Wang, J., Qiu, J., Bui, T., Shrivastava, A., Wang, Z.: Align and attend: Multimodal summarization with dual contrastive losses. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2023)

    2023

  11. [19]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

    Hu, K., Gao, F., Nie, X., Zhou, P., Tran, S., Neiman, T., Wang, L., Shah, M., Hamid, R., Yin, B., Chilimbi, T.: M-llm based video frame selection for efficient video understanding. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). pp...

    2025

  12. [20]

    In: Walsh, T., Shah, J., Kolter, Z

    Hua, H., Tang, Y., Xu, C., Luo, J.: V2xum-llm: Cross-modal video summarization with temporal prompt instruction tuning. In: Walsh, T., Shah, J., Kolter, Z. (eds.) AAAI-25, Sponsored by the Association for the Advancement of Artificial Intelli- gence, February 25 - March 4, 202...

    2025

  13. [21]

    Expert Syst

    Huang, Z., Long, R., Chen, H., Wu, M., Li, Q., Na, J.: LLMQR: efficient knowledge distillation for questionnaire recommendation via large language models. Expert Syst. Appl.303, 130632 (2026). ,

    2026

  14. [22]

    In: Bengio, Y., LeCun, Y

    Kingma, D.P., Ba, J.: Adam: A method for stochastic optimization. In: Bengio, Y., LeCun, Y. (eds.) 3rd International Conference on Learning Representations, ICLR 2015, San Diego, CA, USA, May 7-9, 2015, Conference Track Proceedings (2015),

    2015

  15. [23]

    In: Findings of the Association for Computational Linguistics: EACL 2024

    Konen, K., Jentzsch, S., Diallo, D., Schütt, P., Bensch, O., El Baff, R., Opitz, D., Hecking, T.: Style vectors for steering generative large language models. In: Findings of the Association for Computational Linguistics: EACL 2024. pp. 782–

    2024

  16. [24]

    In: IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2025, Nashville, TN, USA, June 11-15, 2025

    Lee, M.J., Gong, D., Cho, M.: Video summarization with large language models. In: IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2025, Nashville, TN, USA, June 11-15, 2025. pp. 18981–18991. Computer Vision Foundation / IEEE (2025). ,

    2025

  17. [25]

    In: Advances in Neural Information Processing Systems

    Lei, J., Berg, T.L., Bansal, M.: Detecting moments and highlights in videos via natural language queries. In: Advances in Neural Information Processing Systems. vol. 34 (2021),

    2021

  18. [26]

    Li, F., Zhang, R., Zhang, H., Zhang, Y., Li, B., Li, W., Ma, Z., Li, C.: Llava- next-interleave: Tackling multi-image, video, and 3d in large multimodal models (2024),

    2024

  19. [27]

    IEEE Access14, 662–678 (2026)

    Liao, P., Lee, H., Wang, H.: Cross-attention reprogramming for ASR: bridging discrete speech units and pretrained language models. IEEE Access14, 662–678 (2026). ,

    2026

  20. [28]

    IEEE Transactions on Multi- media (2024)

    Lin, J., Hua, H., Chen, M., Li, Y., Hsiao, J., Ho, C., Luo, J.: Videoxum: Cross- modal visual and textural summarization of videos. IEEE Transactions on Multi- media (2024)

    2024

  21. [29]

    Chu and Y

    Liu, D., Whitehouse, C., Yu, X., Mahon, L., Saxena, R., Zhao, Z., Qiu, Y., Lapata, M., Demberg, V.: What is that talk about? A video-to-text summarization dataset 18 R. Chu and Y. Lao for scientific presentations. In: Che, W., Nabende, J., Shutova, E., Pilehvar, M.T. (eds.) Pr...

    2025

  22. [30]

    Ma, F., Jin, X., Wang, H., Xian, Y., Feng, J., Yang, Y.: Vista-llama: Reducing hal- lucinationinvideolanguagemodelsviaequaldistancetovisualtokens.In:Proceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). pp. 13151–13160 (2024)

    2024

  23. [31]

    In: Ku, L.W., Martins, A., Srikumar, V

    Maaz, M., Rasheed, H., Khan, S., Khan, F.: Video-ChatGPT: Towards detailed video understanding via large vision and language models. In: Ku, L.W., Martins, A., Srikumar, V. (eds.) Proceedings of the 62nd Annual Meeting of the Associ- ation for Computational Linguistics (Volume...

    2024

  24. [32]

    In: IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2025, Nashville, TN, USA, June 11-15, 2025

    Meng, N., Manicke, C., Sahu, R., Ding, C., Lao, Y.: Advancing adversarial robust- ness in gnerfs: The il2-nerf attack. In: IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2025, Nashville, TN, USA, June 11-15, 2025. pp. 16388–16397. Computer Vision Foundati...

    2025

  25. [33]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

    Moon, W., Hyun, S., Park, S., Park, D., Heo, J.P.: Query-dependent video rep- resentation for moment retrieval and highlight detection. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). pp. 23023–23033 (June 2023),

    2023

  26. [34]

    In: Ranzato, M., Beygelzimer, A., Dauphin, Y.N., Liang, P., Vaughan, J.W

    Narasimhan, M., Rohrbach, A., Darrell, T.: Clip-it! language-guided video sum- marization. In: Ranzato, M., Beygelzimer, A., Dauphin, Y.N., Liang, P., Vaughan, J.W. (eds.) Advances in Neural Information Processing Systems 34: Annual Con- ference on Neural Information Processin...

    2021

  27. [35]

    Journal of Imaging10(9), 229 (2024)

    Pang, Z., Nakashima, Y., Otani, M., Nagahara, H.: Unleashing the power of con- trastive learning for zero-shot video summarization. Journal of Imaging10(9), 229 (2024)

    2024

  28. [36]

    IEEE Access14, 24082–24095 (2026)

    Park, S., Chung, K.: Spinout: Enhanced rotation-based quantization for LLM by outlier injection. IEEE Access14, 24082–24095 (2026). ,

    2026

  29. [37]

    In: Proceedings of the IEEE International Conference on Computer Vision (ICCV)

    Selvaraju, R.R., Cogswell, M., Das, A., Vedantam, R., Parikh, D., Batra, D.: Grad- cam: Visual explanations from deep networks via gradient-based localization. In: Proceedings of the IEEE International Conference on Computer Vision (ICCV). pp. 618–626 (2017)

    2017

  30. [38]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2024)

    Son, J., Park, J., Kim, K.: Csta: Cnn-based spatiotemporal attention for video summarization. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2024)

    2024

  31. [39]

    In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR)

    Song, Y., Vallmitjana, J., Stent, A., Jaimes, A.: Tvsum: Summarizing web videos using titles. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). pp. 5179–5187 (2015)

    2015

  32. [40]

    In: Interna- tional Conference on Learning Representations (ICLR) (2025),

    Stolfo, A., Balachandran, V., Yousefi, S., Horvitz, E., Nushi, B.: Improving instruction-following in language models through activation steering. In: Interna- tional Conference on Learning Representations (ICLR) (2025),

    2025

  33. [41]

    In: Ghahramani, Z., Welling, M., Cortes, C., Lawrence, N.D., Wein- berger, K.Q

    Sutskever, I., Vinyals, O., Le, Q.V.: Sequence to sequence learning with neural networks. In: Ghahramani, Z., Welling, M., Cortes, C., Lawrence, N.D., Wein- berger, K.Q. (eds.) Advances in Neural Information Processing Systems 27: Annual Conference on Neural Information Proces...

    2014

  34. [42]

    In: Proceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition (CVPR)

    Tang, F., Liu, C., Xu, Z., Hu, M., Huang, Z., Xue, H., Chen, Z., Peng, Z., Yang, Z., Zhou, S., Li, W., Li, Y., Song, W., Su, S., Feng, W., Su, J., Lin, M., Peng, Y., Cheng, X., Razzak, I., Ge, Z.: Seeing far and clearly: Mitigating hallucinations in mllms with attention causal...

    2025

  35. [43]

    AI Soc.41(1), 5–21 (2026)

    de la Torre, P.G., Pérez-Verdugo, M., Barandiaran, X.E.: Attention is all they need: cognitive science and the (techno)political economy of attention in humans and machines. AI Soc.41(1), 5–21 (2026). ,

    2026

  36. [44]

    In: Guyon, I., von Luxburg, U., Bengio, S., Wallach, H.M., Fergus, R., Vishwanathan, S.V.N., Garnett, R

    Vaswani,A.,Shazeer,N.,Parmar,N.,Uszkoreit,J.,Jones,L.,Gomez,A.N.,Kaiser, L., Polosukhin, I.: Attention is all you need. In: Guyon, I., von Luxburg, U., Bengio, S., Wallach, H.M., Fergus, R., Vishwanathan, S.V.N., Garnett, R. (eds.) Advances in Neural Information Processing Sys...

    2017

  37. [45]

    In: Proceedings of the 43rd IEEE/ACM International Conference on Computer-Aided Design

    Wang, A., Zhang, K., Parhi, K.K., Lao, Y.: Hermes: Homomorphic encryption over residual number system for multi-level evaluations. In: Proceedings of the 43rd IEEE/ACM International Conference on Computer-Aided Design. ICCAD ’24, Association for Computing Machinery, New York, ...

    2025

  38. [46]

    In: 2016 IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2016, Las Vegas, NV, USA, June 27-30, 2016

    Yao, T., Mei, T., Rui, Y.: Highlight detection with pairwise deep ranking for first- person video summarization. In: 2016 IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2016, Las Vegas, NV, USA, June 27-30, 2016. pp. 982–990. IEEE Computer Society (2016). ,

    2016

  39. [47]

    Ye, J., Xu, H., Liu, H., Hu, A., Yan, M., Qian, Q., Zhang, J., Huang, F., Zhou, J.: mplug-owl3: Towards long image-sequence understanding in multi-modal large language models (2024),

    2024

  40. [48]

    IEEE CAA J

    Yu, Q., Wang, Z., Wei, G., Yu, H.: Deep learning for video summarization: Sys- tematic review, challenges and opportunities. IEEE CAA J. Autom. Sinica13(1), 21–42 (2026). ,

    2026

  41. [49]

    Displays91, 103269 (2026)

    Zhang, H., Shi, P., Yang, G.: Tdtoon: Two-stage diffusion for controllable cartoon video generation via sketch enhancement. Displays91, 103269 (2026). ,

    2026

  42. [50]

    In: 2024 58th Asilomar Conference on Signals, Systems, and Computers

    Zhang, K., Wang, A., Parhi, K.K., Lao, Y.: Hardware acceleration for fully homo- morphic encryption scheme switching from ckks to fhew. In: 2024 58th Asilomar Conference on Signals, Systems, and Computers. pp. 1792–1796 (2024)

    2024

  43. [51]

    Zhao, C., Ding, G., Wang, W., Yang, Z., Liu, Z., Chen, H., Shen, C.: Freercustom: Training-free multi-concept customization for image and video generation. Int. J. Comput. Vis.134(1), 17 (2026). ,

    2026

  44. [52]

    IEEE Transactions on Image Processing30, 948–962 (2021)

    Zhu, W., Lu, J., Li, J., Zhou, J.: Dsnet: A flexible detect-to-summarize network for video summarization. IEEE Transactions on Image Processing30, 948–962 (2021)

    2021

  45. [802]

    Julian’s, Malta (March 2024)

    Association for Computational Linguistics, St. Julian’s, Malta (March 2024). ,

    2024

Pith tools

Reviewed August 1, 2026 · model on record in the stance chip above.