Molmo2 delivers state-of-the-art open-weight video VLMs with new grounding datasets and training methods that outperform prior open models and match or exceed some proprietary ones on pointing and tracking tasks.
super hub Baseline reference
The Kinetics Human Action Video Dataset
Baseline reference. 62% of citing Pith papers use this work as a benchmark or comparison.
119
Pith papers citing it
Baseline
62% of classified citations
abstract
We describe the DeepMind Kinetics human action video dataset. The dataset contains 400 human action classes, with at least 400 video clips for each action. Each clip lasts around 10s and is taken from a different YouTube video. The actions are human focussed and cover a broad range of classes including human-object interactions such as playing instruments, as well as human-human interactions such as shaking hands. We describe the statistics of the dataset, how it was collected, and give some baseline performance figures for neural network architectures trained and tested for human action classification on this dataset. We also carry out a preliminary analysis of whether imbalance in the dataset leads to bias in the classifiers.
hub tools
citation-role summary
dataset 17
background 9
citation-polarity summary
claims ledger
- abstract We describe the DeepMind Kinetics human action video dataset. The dataset contains 400 human action classes, with at least 400 video clips for each action. Each clip lasts around 10s and is taken from a different YouTube video. The actions are human focussed and cover a broad range of classes including human-object interactions such as playing instruments, as well as human-human interactions such as shaking hands. We describe the statistics of the dataset, how it was collected, and give some baseline performance figures for neural network architectures trained and tested for human action class
authors
co-cited works
representative citing papers
Introduces the BAH dataset with 1,427 annotated videos for multimodal recognition of ambivalence/hesitancy in digital behavior change contexts.
MMDG-Bench provides unified protocols and ten baselines for multimodal domain generalization, showing structured DG-MML combinations often outperform prior methods with insights on framework choice and backbone effects.
VideoABC estimates video-LLM failure probability via low-dimensional attribute projection, dual quantization (k-means plus lattice), and psychophysics-inspired synthetic data.
Domain-incremental video learning that permits forgetting through per-domain LoRA adapters and recovers the matching adapter at inference via test-time training on a self-supervised MAE reconstruction head.
YoCausal benchmark shows video diffusion models detect the arrow of time but lack genuine causal understanding relative to humans.
Introduces the USV dataset of 224K short user-generated videos and benchmarks topic recognition plus video-text retrieval with MMF-Net and VTCL baselines.
Introduces FogAct paired clean-foggy video dataset and FogNet two-stream CLIP model that learns fog-invariant semantic representations via clean-video guidance.
PEIRA learns predictive encoders by optimizing the trace of the optimal inter-view linear regressor, with only nontrivial global minimizers as stable equilibria that recover leading nonlinear canonical correlation subspaces.
Minerva-Ego is a new benchmark for egocentric visual reasoning with dense human-annotated traces and masks, showing that spatiotemporal hints substantially improve frontier model performance.
PoseBridge recovers semantic information lost during skeletonization by extracting pose-anchored cues from human pose estimation and transferring them via skeleton-conditioned bridging and semantic prototype adaptation, yielding 13.3-17.4 point gains on the Kinetics PURLS benchmark.
RaPO reduces catastrophic forgetting in visual continual learning by shaping rewards around policy drift and stabilizing advantages with cross-task exponential moving averages during reinforcement fine-tuning of multimodal models.
LMMs perceive videos but underexploit visual content for causal reasoning due to textual shortcuts; ProCauEval diagnoses this and ADPO training reduces reliance on priors.
EyeCue detects driver cognitive distraction by modeling gaze-visual context interactions in egocentric videos and achieves 74.38% accuracy on the new CogDrive dataset, outperforming 11 baselines.
Temporal information in Video-LLMs is encoded well by video-centric encoders but disrupted by standard projectors; time-preserved MLPs plus AoT supervision yield 98.1% accuracy on arrow-of-time and gains on other temporal tasks.
McNdroid is a new longitudinal multimodal benchmark showing that Android malware detectors degrade over time but multimodal approaches maintain better performance across long temporal gaps.
SIGMA-ASL is a multimodal dataset with 93,545 word-level ASL clips from Kinect RGB-D, mmWave radar, and dual IMUs, plus benchmarking protocols for single- and multi-modal recognition.
VEBENCH is the first benchmark with 3.9K videos and 3,080 human-verified QA pairs that measures LMMs on video editing technique recognition and operation simulation, revealing a large gap to human performance.
SignMAE uses segmentation-driven masking in a mask-and-reconstruct self-supervised task to learn fine-grained sign representations, achieving state-of-the-art accuracy on WLASL, NMFs-CSL, and Slovo with fewer frames and modalities.
VAnim creates open-domain text-to-SVG animations via sparse state updates on a persistent DOM tree, identification-first planning, and rendering-aware RL with a new 134k-example benchmark.
RefVQA uses a query-centered reference graph and graph-guided difference aggregation to improve AI-generated video quality assessment by incorporating inter-video comparisons.
GTASA supplies annotated multi-actor videos with exact 3D spatial and temporal ground truth that outperforms neural video generators in physical and semantic validity while enabling new probes of video encoders.
LMFT enables state-of-the-art performance in video unsupervised domain adaptation by focusing on motion-rich tokens and reducing computational overhead.
InstrAction pretrains video foundation models using action-centric data filtering, hard negatives, an Action Perceiver module, DTW-Align, and Masked Action Modeling to reduce static bias and outperform prior models on a new InstrAct Bench for semantic, procedural, and retrieval tasks.
citing papers explorer
-
Molmo2: Open Weights and Data for Vision-Language Models with Video Understanding and Grounding
Molmo2 delivers state-of-the-art open-weight video VLMs with new grounding datasets and training methods that outperform prior open models and match or exceed some proprietary ones on pointing and tracking tasks.
-
BAH Dataset for Ambivalence/Hesitancy Recognition in Videos for Digital Behavioural Change
Introduces the BAH dataset with 1,427 annotated videos for multimodal recognition of ambivalence/hesitancy in digital behavior change contexts.
-
MMDG-Bench: A Benchmark for Multimodal Domain Generalization
MMDG-Bench provides unified protocols and ten baselines for multimodal domain generalization, showing structured DG-MML combinations often outperform prior methods with insights on framework choice and backbone effects.
-
An Attribute-Based Measure of Video Complexity
VideoABC estimates video-LLM failure probability via low-dimensional attribute projection, dual quantization (k-means plus lattice), and psychophysics-inspired synthetic data.
-
Remembering by Reconstructing: Domain Incremental Learning With Test-Time Training on Video Streams
Domain-incremental video learning that permits forgetting through per-domain LoRA adapters and recovers the matching adapter at inference via test-time training on a self-supervised MAE reconstruction head.
-
YoCausal: How Far is Video Generation from World Model? A Causality Perspective
YoCausal benchmark shows video diffusion models detect the arrow of time but lack genuine causal understanding relative to humans.
-
USV: Towards Understanding the User-generated Short-form Videos
Introduces the USV dataset of 224K short user-generated videos and benchmarks topic recognition plus video-text retrieval with MMF-Net and VTCL baselines.
-
Seeing Through Fog: Towards Fog-Invariant Action Recognition
Introduces FogAct paired clean-foggy video dataset and FogNet two-stream CLIP model that learns fog-invariant semantic representations via clean-video guidance.
-
PEIRA: Learning Predictive Encoders through Inter-View Regressor Alignment
PEIRA learns predictive encoders by optimizing the trace of the optimal inter-view linear regressor, with only nontrivial global minimizers as stable equilibria that recover leading nonlinear canonical correlation subspaces.
-
Minerva-Ego: Spatiotemporal Hints for Egocentric Video Understanding
Minerva-Ego is a new benchmark for egocentric visual reasoning with dense human-annotated traces and masks, showing that spatiotemporal hints substantially improve frontier model performance.
-
PoseBridge: Bridging the Skeletonization Gap for Zero-Shot Skeleton-Based Action Recognition
PoseBridge recovers semantic information lost during skeletonization by extracting pose-anchored cues from human pose estimation and transferring them via skeleton-conditioned bridging and semantic prototype adaptation, yielding 13.3-17.4 point gains on the Kinetics PURLS benchmark.
-
Overcoming Catastrophic Forgetting in Visual Continual Learning with Reinforcement Fine-Tuning
RaPO reduces catastrophic forgetting in visual continual learning by shaping rewards around policy drift and stabilizing advantages with cross-task exponential moving averages during reinforcement fine-tuning of multimodal models.
-
Perception Without Engagement: Dissecting the Causal Discovery Deficit in LMMs
LMMs perceive videos but underexploit visual content for causal reasoning due to textual shortcuts; ProCauEval diagnoses this and ADPO training reduces reliance on priors.
-
EyeCue: Driver Cognitive Distraction Detection via Gaze-Empowered Egocentric Video Understanding
EyeCue detects driver cognitive distraction by modeling gaze-visual context interactions in egocentric videos and achieves 74.38% accuracy on the new CogDrive dataset, outperforming 11 baselines.
-
Tracing the Arrow of Time: Diagnosing Temporal Information Flow in Video-LLMs
Temporal information in Video-LLMs is encoded well by video-centric encoders but disrupted by standard projectors; time-preserved MLPs plus AoT supervision yield 98.1% accuracy on arrow-of-time and gains on other temporal tasks.
-
McNdroid: A Longitudinal Multimodal Benchmark for Robust Drift Detection in Android Malware
McNdroid is a new longitudinal multimodal benchmark showing that Android malware detectors degrade over time but multimodal approaches maintain better performance across long temporal gaps.
-
SIGMA-ASL: Sensor-Integrated Multimodal Dataset for Sign Language Recognition
SIGMA-ASL is a multimodal dataset with 93,545 word-level ASL clips from Kinect RGB-D, mmWave radar, and dual IMUs, plus benchmarking protocols for single- and multi-modal recognition.
-
VEBench:Benchmarking Large Multimodal Models for Real-World Video Editing
VEBENCH is the first benchmark with 3.9K videos and 3,080 human-verified QA pairs that measures LMMs on video editing technique recognition and operation simulation, revealing a large gap to human performance.
-
SignMAE: Segmentation-Driven Self-Supervised Learning for Sign Language Recognition
SignMAE uses segmentation-driven masking in a mask-and-reconstruct self-supervised task to learn fine-grained sign representations, achieving state-of-the-art accuracy on WLASL, NMFs-CSL, and Slovo with fewer frames and modalities.
-
VAnim: Rendering-Aware Sparse State Modeling for Structure-Preserving Vector Animation
VAnim creates open-domain text-to-SVG animations via sparse state updates on a persistent DOM tree, identification-first planning, and rendering-aware RL with a new 134k-example benchmark.
-
Comparison Drives Preference: Reference-Aware Modeling for AI-Generated Video Quality Assessment
RefVQA uses a query-centered reference graph and graph-guided difference aggregation to improve AI-generated video quality assessment by incorporating inter-video comparisons.
-
GTASA: Ground Truth Annotations for Spatiotemporal Analysis, Evaluation and Training of Video Models
GTASA supplies annotated multi-actor videos with exact 3D spatial and temporal ground truth that outperforms neural video generators in physical and semantic validity while enabling new probes of video encoders.
-
Learnable Motion-Focused Tokenization for Effective and Efficient Video Unsupervised Domain Adaptation
LMFT enables state-of-the-art performance in video unsupervised domain adaptation by focusing on motion-rich tokens and reducing computational overhead.
-
InstrAct: Towards Action-Centric Understanding in Instructional Videos
InstrAction pretrains video foundation models using action-centric data filtering, hard negatives, an Action Perceiver module, DTW-Align, and Masked Action Modeling to reduce static bias and outperform prior models on a new InstrAct Bench for semantic, procedural, and retrieval tasks.
-
InstAP: Instance-Aware Vision-Language Pre-Train for Spatial-Temporal Understanding
InstAP introduces instance-aware pre-training with a new dual-granularity dataset InstVL that improves both fine-grained instance retrieval and global video understanding over standard VLP baselines.
-
MotionScape: A Large-Scale Real-World Highly Dynamic UAV Video Dataset for World Models
MotionScape is a large-scale UAV video dataset with highly dynamic 6-DoF motions, geometric trajectories, and semantic annotations to train world models that better simulate complex 3D dynamics under large viewpoint changes.
-
Recurrent Video Masked Autoencoders
RVM uses recurrent computation inside a masked autoencoder to learn video representations that match or exceed prior video and image models on classification, tracking, and dense spatial tasks with up to 30x better parameter efficiency.
-
HSG-12M: A Large-Scale Benchmark of Spatial Multigraphs from the Energy Spectra of Non-Hermitian Crystals
HSG-12M is a large dataset of spatial multigraphs derived from non-Hermitian crystal energy spectra via the Poly2Graph pipeline, positioned as the first large-scale benchmark of this graph type.
-
SIV-Bench: A Video Benchmark for Social Interaction Understanding and Reasoning
SIV-Bench is a new video benchmark with 2,792 clips and 5,455 QA pairs that evaluates MLLMs on social scene understanding, state reasoning, and dynamics prediction using social relation theory.
-
History-Guided Video Diffusion
DFoT enables flexible history conditioning in video diffusion, with history guidance methods that boost temporal consistency and support long rollouts.
-
MLVU: Benchmarking Multi-task Long Video Understanding
MLVU is a new benchmark for long video understanding that uses extended videos across diverse genres and multi-task evaluations, revealing that current MLLMs struggle significantly and degrade sharply with longer durations.
-
Language Model Beats Diffusion -- Tokenizer is Key to Visual Generation
A new shared video-image tokenizer enables large language models to surpass diffusion models on standard visual generation benchmarks.
-
Video Diffusion Models
A diffusion model for video generation extends image architectures with joint image-video training and improved conditional sampling, delivering first large-scale text-to-video results and state-of-the-art performance on video prediction and unconditional generation benchmarks.
-
Switchable Normalization for Learning-to-Normalize Deep Representation
Switchable Normalization learns per-layer weights to combine channel, layer, and minibatch normalizers, claiming robustness to batch size and better results than fixed normalizers on ImageNet, COCO, CityScapes, ADE20K, MegaFace, and Kinetics.
-
SVI-Bench: A Dynamic Microworld for Strategic Video Intelligence
SVI-Bench is a 35K-hour sports video benchmark with 9 tasks across four cognitive pillars that reveals multimodal models drop from ~73% on action QA to 5% on agentic evidence-gathering tasks.
-
TeachObs: A Human-Validated Benchmark for Multimodal Teaching Observation and Model Evaluation
TeachObs is a new human-validated benchmark dataset and evaluation protocol for multimodal AI on classroom teaching observation, showing no model dominates across tracks and that models over-rate procedurally clear lessons.
-
Mining Multi-Modality Spatio-Temporal Cues for Video Important Person Identification
Introduces VIP identification task, releases Temporal-VIP dataset, and presents VIP-Net framework that achieves 67.3% accuracy on identifying important persons in videos while providing rationale similarity of 0.63.
-
Cambrian-P: Pose-Grounded Video Understanding
Cambrian-P adds per-frame camera pose tokens and a regression head to video MLLMs, delivering 4.5-6.5% gains on spatial benchmarks, generalization to other video QA tasks, and SOTA streaming pose estimation on ScanNet.
-
SpecSem-Net: Integrating Spectral and Semantic Features for Robust AI-generated Video Detection
SpecSem-Net integrates Fourier-based spectral filtering with semantic-guided gated merging to detect AI-generated videos, reporting 87.25% accuracy on a new benchmark of five commercial generators and 95.59% on public datasets.
-
Bridging Brain and Semantics: A Hierarchical Framework for Semantically Enhanced fMRI-to-Video Reconstruction
CineNeuron improves fMRI-to-video reconstruction by combining bottom-up semantic enrichment with top-down Mixture-of-Memories integration and outperforms prior methods on benchmarks.
-
HumanNet: Scaling Human-centric Video Learning to One Million Hours
HumanNet is a 1M-hour human-centric video dataset with interaction annotations that enables better vision-language-action model performance than equivalent robot data in a controlled test.
-
Detecting AI-Generated Videos with Spiking Neural Networks
MAST with spiking neural networks achieves 93.14% mean accuracy detecting AI-generated videos from 10 unseen generators by exploiting smoother pixel residuals and compact semantic trajectories.
-
Multimodal Learning on Low-Quality Data with Conformal Predictive Self-Calibration
CPSC uses conformal prediction to decompose and fuse robust unimodal features and recalibrate gradients based on instance reliability, outperforming prior methods on imbalanced and noisy multimodal benchmarks.
-
Featurising Pixels from Dynamic 3D Scenes with Linear In-Context Learners
LILA learns temporally consistent semantic and geometric pixel features from uncurated videos via linear in-context learning on off-the-shelf depth and motion cues, yielding empirical gains on video object segmentation, surface normal estimation, and semantic segmentation.
-
$\text{PKS}^4$:Parallel Kinematic Selective State Space Scanners for Efficient Video Understanding
PKS^4 adds a kinematic-prior-driven parallel state space scanner module to 2D vision backbones for linear-complexity temporal modeling in videos, delivering SOTA action recognition with 10x lower training compute and convergence in 20 epochs.
-
Exploring High-Order Self-Similarity for Video Understanding
The MOSS module learns and combines multi-order space-time self-similarity features to enhance temporal dynamics modeling in videos across action recognition, VQA, and robotic tasks.
-
Multi-modal Test-time Adaptation via Adaptive Probabilistic Gaussian Calibration
A probabilistic Gaussian model with adaptive contrastive asymmetry rectification improves multi-modal test-time adaptation by modeling category distributions and correcting modality asymmetry for better predictions under shifts.
-
EAST: Early Action Prediction Sampling Strategy with Token Masking
EAST uses randomized time-step sampling and token masking to train a single encoder-only model that generalizes across all observation ratios in early action prediction and reports new state-of-the-art accuracy on NTU60, SSv2, and UCF101.
-
Identifying Ethical Biases in Action Recognition Models
The authors create a synthetic video auditing framework that detects statistically significant skin color biases in popular human action recognition models even when actions are identical.
-
One Token per Highly Selective Frame: Towards Extreme Compression for Long Video Understanding
XComp reaches extreme video compression (one token per selective frame) via learnable progressive token compression and question-conditioned frame selection, lifting LVBench accuracy from 42.9 percent to 46.2 percent after tuning on 2.5 percent of standard data.