Papers

Field

Top tools here

Open problems in Multimodal

Lack of holistic video understanding benchmarks1 paper raises it · 1 addresses it
Existing benchmarks cover only partial evaluation dimensions, limiting systematic diagnosis of omni-modal video understanding.
- OmniCompass: A Holistic Benchmark for Omni-Modal Video Grounding and Understandingraises · addresses
Long-horizon task execution and recovery1 paper raises it · 0 address it
Agents struggle with preserving progress, recovering from failures, and verifying physical states over long sequences.
- Has GPT-6-Astra Achieved Embodied Intelligence?raises
Efficient spatial reasoning for real-time interaction2 papers raise it · 2 address it
Current models use slow autoregressive decoding for compact output spaces, limiting interaction speed.
- VideoJev: System One Model for Video Understanding and Interactionraises · addresses
- MJev-Spatial: System One Spatial Intelligence via Native Attention Matchingraises · addresses
Cross-shot consistency in agentic video generation1 paper raises it · 1 addresses it
Maintaining spatial and temporal consistency across non-consecutive video shots is difficult.
- Any Time, Any Space: A World-State-Centric Framework for Agentic Video Generationraises · addresses
Precise grounding for physical intelligence1 paper raises it · 1 addresses it

Open problems

Recently funded

Topics in this field

video understanding · video language · video generation301 papers50 this week
visual question answering · vision language models · vision language165 papers46 this week
autoencoders · imagenet · autoencoder224 papers43 this week
video generation · efficient video · autoregressive video181 papers32 this week
visual reasoning · multimodal reasoning · visual question answering177 papers32 this week
semantic segmentation · vision language · attention202 papers31 this week

Topics in this field

Active contributors · 90 days

Active contributors

Activity · 12 weeks

Coming with more sources: repos

Activity

Recent papers

All in Research →

arXiv cs.CVOct 16
PONDER: Learning to Generate in the Right Order
Julius Erbach, Dominik Narnhofer, Bernt Schiele, Jan Eric Lenssen, Konrad Schindler
We introduce PONDER, a method that learns in which order an image generator should fill in content and which of its earlier predictions to revise. Masked generative models construct a sample by repeatedly filling in subsets of missing variables. The order of these partial predictions determines the context available in subsequent steps and therefore affects the quality of the final sample. Common strategies are to choose the order randomly or based on local confidence estimates. Conditioned on the partially generated sample and the generator's features, our policy chooses which positions to fill next and/or which previously filled positions to mask for later regeneration, a mechanism we call rethinking. In this way, the generation process is able to adapt to the evolving context and also to revise earlier predictions. We optimize the policy using group-relative policy gradients, with rewards that measure constraint satisfaction, human preference, or distribution quality. For the latter, we introduce FIDgrad, an image-level reward derived from the (first-order) change in Fréchet Inception Distance caused by adding a generated sample to the current feature distribution. With the…
PDF arXiv
arXiv cs.CVOct 16
When Evidence Vanishes: Distilling Visual-Evidence Dependency in Medical Vision-Language Models
Qizhen Lan, Yiling Ma, Ruiqi Chen, Mengchen Fan, Xi Xiao +6
Compact medical vision-language models (VLMs) can answer clinical questions correctly without reliably using the image, relying instead on question wording or learned priors. Standard knowledge distillation copies a teacher's answer preference on the original image, but it leaves unchecked whether that preference is conditional on the visual evidence required by the question. We study this missing target as medical visual evidence dependency: support for the correct answer should decrease when the matched evidence is absent. We introduce Evidence-Conditioned Dependency Distillation (ECDD), which pairs each training image with a control matched to the evidence level and distills the teacher's drop in answer margin across the pair. A lower bound objective transfers positive dependency by penalizing only shortfalls, while original-image supervision preserves task accuracy. The same formulation applies when the evidence is a queried clinical region, a surgical instrument, or whole-image visual content. Across three medical VQA settings, ECDD keeps original-image accuracy competitive with distillation and alignment baselines while making answer support more sensitive to evidence…
PDF arXiv
arXiv cs.CLOct 16
HydraLLM: Adaptive Width Slicing for Elastic Text and Vision-Language Models
Oscar Kemme, Janek Haberer, Ali Hojjat, Olaf Landsiedel
LLMs and LVLMs are increasingly deployed across heterogeneous hardware, ranging from data-center GPUs to resource-constrained edge devices. Their fixed-size architectures limit the ability to reuse a single model across deployment settings with different compute, latency, and model-size targets. We introduce HydraLLM, an elastic MQA-based architecture. Its number of active attention heads determines the effective widths of the embeddings, attention projections, feed-forward layers, and the output head, forming a nested family of subnetworks within a single parameter set. We introduce a static-to-elastic conversion pipeline that, starting from a pre-trained model checkpoint, includes importance-driven weight reordering and stochastic subnetwork training with knowledge distillation. We extend this principle to multimodal inference through an elastic projection module whose output width follows the active text width. HydraLLM covers a wider range of active parameter configurations than FFN-only elastic architectures. With 75% of heads active, HydraLLM achieves 2.2 percentage points higher TriviaQA accuracy than a state-of-the-art nested baseline configuration which has 17.25% more…
PDF arXiv
arXiv cs.CVOct 16
ShadowCast: On-the-Fly Context-Aware Shadow Augmentation for Remote Sensing Image Segmentation
Yechan Kim, Dongho Yoon, Jongmin Joo, JongHyun Park, Moongu Jeon
Shadows cast by elevated objects alter the spectral appearance and obscure the boundaries of objects in remote sensing images, thereby degrading semantic segmentation performance. Most existing approaches handle shadows as degradations to be detected and removed, requiring additional costs from dedicated networks and shadow-specific annotations. We propose ShadowCast, which takes the opposite approach: it synthesizes shadows and injects them during training for remote sensing segmentation. It extracts masks of elevated objects, such as buildings and trees, directly from the segmentation labels and estimates an image-specific cast-shadow direction cue by probing the brightness around them. It then projects synthetic shadows along the inferred cast-shadow direction, requiring no learned network and no acquisition metadata. Experiments on three widely used remote sensing benchmarks, Vaihingen, Potsdam, and LoveDA, show that ShadowCast consistently outperforms the baseline across diverse state-of-the-art segmentation architectures. Moreover, it brings additional performance gains when combined with standard augmentation and regularization techniques, indicating that stochastic shadow…
PDF arXiv
arXiv cs.AIOct 16
Generative Policy Improvement from Observational Data
Jonas Schweisthal, Dennis Frauen, Stefan Feuerriegel, Emma Brunskill
We study how to improve decision policies from observational data to generate better actions when actions are complex, structured objects, such as text messages or images. This task is subject to two major challenges: (i) action-level overlap is often too weak for reliable off-policy evaluation and improvement, and (ii) generating new structured actions often requires expensive training or fine-tuning (e.g., of an LLM) and can be ill-defined due to information loss in action representations. To address both, we propose PolicyLift, a three-stage framework for generative policy improvement with structured actions from observational data. In Stage 1, PolicyLift learns a small set of discrete action codes that capture high-level strategies in the observed action space and are tailored to support policy improvement under limited overlap. In Stage 2, it learns a context-dependent policy over these action codes. At deployment time, Stage 3 "lifts" the learned code-level policy back to the original action space by generating actions from the historical behavior system in a code-conditioned way. We show theoretically that improving the policy over action codes guarantees improvement of…
PDF arXiv
arXiv cs.CVOct 16
Momentum: Persistent Dynamics for Video World Models
Nedko Savov, Stefan Maria Ailuro, Danda Pani Paudel, Yilun Du, Xi Wang +1
Video world models must capture how a scene evolves beyond the current camera view to generate consistent observations when its content reappears. Existing approaches rely on observation history, static point clouds, or image caches constructed from past frames. These representations are often expensive and, more importantly, cannot update dynamic scene content outside the field of view: generated vehicles and pedestrians may be forgotten after leaving the field of view or reappear in physically inaccurate states. We represent an evolving scene through synchronized videos from reference-camera rigs, each comprising eight cameras that jointly cover the surrounding scene. We connect these rigs in a graph and propagate video generation between neighboring rigs to coordinate scene dynamics across locations and time. We introduce Momentum, a two-stage pipeline: WorldUpdate jointly generates reference-camera videos as a shared memory of scene dynamics, and WorldInteract grounds camera-controlled video generation in this evolving memory, allowing moving content to remain represented outside the target view and reappear consistently. To support training and evaluation, we develop a CARLA…
PDF arXiv

Papers

Topics in this field

video understanding · video language · video generation301 papers50 this week
visual question answering · vision language models · vision language165 papers46 this week
autoencoders · imagenet · autoencoder224 papers43 this week
video generation · efficient video · autoregressive video181 papers32 this week
visual reasoning · multimodal reasoning · visual question answering177 papers32 this week
semantic segmentation · vision language · attention202 papers31 this week

Active contributors · 90 days

Activity · 12 weeks

Coming with more sources: repos

Recent papers

All in Research →

arXiv cs.CVOct 16
PONDER: Learning to Generate in the Right Order
Julius Erbach, Dominik Narnhofer, Bernt Schiele, Jan Eric Lenssen, Konrad Schindler
We introduce PONDER, a method that learns in which order an image generator should fill in content and which of its earlier predictions to revise. Masked generative models construct a sample by repeatedly filling in subsets of missing variables. The order of these partial predictions determines the context available in subsequent steps and therefore affects the quality of the final sample. Common strategies are to choose the order randomly or based on local confidence estimates. Conditioned on the partially generated sample and the generator's features, our policy chooses which positions to fill next and/or which previously filled positions to mask for later regeneration, a mechanism we call rethinking. In this way, the generation process is able to adapt to the evolving context and also to revise earlier predictions. We optimize the policy using group-relative policy gradients, with rewards that measure constraint satisfaction, human preference, or distribution quality. For the latter, we introduce FIDgrad, an image-level reward derived from the (first-order) change in Fréchet Inception Distance caused by adding a generated sample to the current feature distribution. With the…
PDF arXiv
arXiv cs.CVOct 16
When Evidence Vanishes: Distilling Visual-Evidence Dependency in Medical Vision-Language Models
Qizhen Lan, Yiling Ma, Ruiqi Chen, Mengchen Fan, Xi Xiao +6
Compact medical vision-language models (VLMs) can answer clinical questions correctly without reliably using the image, relying instead on question wording or learned priors. Standard knowledge distillation copies a teacher's answer preference on the original image, but it leaves unchecked whether that preference is conditional on the visual evidence required by the question. We study this missing target as medical visual evidence dependency: support for the correct answer should decrease when the matched evidence is absent. We introduce Evidence-Conditioned Dependency Distillation (ECDD), which pairs each training image with a control matched to the evidence level and distills the teacher's drop in answer margin across the pair. A lower bound objective transfers positive dependency by penalizing only shortfalls, while original-image supervision preserves task accuracy. The same formulation applies when the evidence is a queried clinical region, a surgical instrument, or whole-image visual content. Across three medical VQA settings, ECDD keeps original-image accuracy competitive with distillation and alignment baselines while making answer support more sensitive to evidence…
PDF arXiv
arXiv cs.CLOct 16
HydraLLM: Adaptive Width Slicing for Elastic Text and Vision-Language Models
Oscar Kemme, Janek Haberer, Ali Hojjat, Olaf Landsiedel
LLMs and LVLMs are increasingly deployed across heterogeneous hardware, ranging from data-center GPUs to resource-constrained edge devices. Their fixed-size architectures limit the ability to reuse a single model across deployment settings with different compute, latency, and model-size targets. We introduce HydraLLM, an elastic MQA-based architecture. Its number of active attention heads determines the effective widths of the embeddings, attention projections, feed-forward layers, and the output head, forming a nested family of subnetworks within a single parameter set. We introduce a static-to-elastic conversion pipeline that, starting from a pre-trained model checkpoint, includes importance-driven weight reordering and stochastic subnetwork training with knowledge distillation. We extend this principle to multimodal inference through an elastic projection module whose output width follows the active text width. HydraLLM covers a wider range of active parameter configurations than FFN-only elastic architectures. With 75% of heads active, HydraLLM achieves 2.2 percentage points higher TriviaQA accuracy than a state-of-the-art nested baseline configuration which has 17.25% more…
PDF arXiv
arXiv cs.CVOct 16
ShadowCast: On-the-Fly Context-Aware Shadow Augmentation for Remote Sensing Image Segmentation
Yechan Kim, Dongho Yoon, Jongmin Joo, JongHyun Park, Moongu Jeon
Shadows cast by elevated objects alter the spectral appearance and obscure the boundaries of objects in remote sensing images, thereby degrading semantic segmentation performance. Most existing approaches handle shadows as degradations to be detected and removed, requiring additional costs from dedicated networks and shadow-specific annotations. We propose ShadowCast, which takes the opposite approach: it synthesizes shadows and injects them during training for remote sensing segmentation. It extracts masks of elevated objects, such as buildings and trees, directly from the segmentation labels and estimates an image-specific cast-shadow direction cue by probing the brightness around them. It then projects synthetic shadows along the inferred cast-shadow direction, requiring no learned network and no acquisition metadata. Experiments on three widely used remote sensing benchmarks, Vaihingen, Potsdam, and LoveDA, show that ShadowCast consistently outperforms the baseline across diverse state-of-the-art segmentation architectures. Moreover, it brings additional performance gains when combined with standard augmentation and regularization techniques, indicating that stochastic shadow…
PDF arXiv
arXiv cs.AIOct 16
Generative Policy Improvement from Observational Data
Jonas Schweisthal, Dennis Frauen, Stefan Feuerriegel, Emma Brunskill
We study how to improve decision policies from observational data to generate better actions when actions are complex, structured objects, such as text messages or images. This task is subject to two major challenges: (i) action-level overlap is often too weak for reliable off-policy evaluation and improvement, and (ii) generating new structured actions often requires expensive training or fine-tuning (e.g., of an LLM) and can be ill-defined due to information loss in action representations. To address both, we propose PolicyLift, a three-stage framework for generative policy improvement with structured actions from observational data. In Stage 1, PolicyLift learns a small set of discrete action codes that capture high-level strategies in the observed action space and are tailored to support policy improvement under limited overlap. In Stage 2, it learns a context-dependent policy over these action codes. At deployment time, Stage 3 "lifts" the learned code-level policy back to the original action space by generating actions from the historical behavior system in a code-conditioned way. We show theoretically that improving the policy over action codes guarantees improvement of…
PDF arXiv
arXiv cs.CVOct 16
Momentum: Persistent Dynamics for Video World Models
Nedko Savov, Stefan Maria Ailuro, Danda Pani Paudel, Yilun Du, Xi Wang +1
Video world models must capture how a scene evolves beyond the current camera view to generate consistent observations when its content reappears. Existing approaches rely on observation history, static point clouds, or image caches constructed from past frames. These representations are often expensive and, more importantly, cannot update dynamic scene content outside the field of view: generated vehicles and pedestrians may be forgotten after leaving the field of view or reappear in physically inaccurate states. We represent an evolving scene through synchronized videos from reference-camera rigs, each comprising eight cameras that jointly cover the surrounding scene. We connect these rigs in a graph and propagate video generation between neighboring rigs to coordinate scene dynamics across locations and time. We introduce Momentum, a two-stage pipeline: WorldUpdate jointly generates reference-camera videos as a shared memory of scene dynamics, and WorldInteract grounds camera-controlled video generation in this evolving memory, allowing moving content to remain represented outside the target view and reappear consistently. To support training and evaluation, we develop a CARLA…
PDF arXiv

Open problems in Multimodal

Lack of holistic video understanding benchmarks1 paper raises it · 1 addresses it
Existing benchmarks cover only partial evaluation dimensions, limiting systematic diagnosis of omni-modal video understanding.
- OmniCompass: A Holistic Benchmark for Omni-Modal Video Grounding and Understandingraises · addresses
Long-horizon task execution and recovery1 paper raises it · 0 address it
Agents struggle with preserving progress, recovering from failures, and verifying physical states over long sequences.
- Has GPT-6-Astra Achieved Embodied Intelligence?raises
Efficient spatial reasoning for real-time interaction2 papers raise it · 2 address it
Current models use slow autoregressive decoding for compact output spaces, limiting interaction speed.
- VideoJev: System One Model for Video Understanding and Interactionraises · addresses
- MJev-Spatial: System One Spatial Intelligence via Native Attention Matchingraises · addresses
Cross-shot consistency in agentic video generation1 paper raises it · 1 addresses it
Maintaining spatial and temporal consistency across non-consecutive video shots is difficult.
- Any Time, Any Space: A World-State-Centric Framework for Agentic Video Generationraises · addresses
Precise grounding for physical intelligence1 paper raises it · 1 addresses it

General-purpose backbones fail to precisely ground objects in clutter or for tiny targets at control speeds.

GroundingPI: A Grounding Foundation Model towards Physical Intelligence with Visual Primitivesraises · addresses

Named by an AI from these papers' abstracts.

Fields · AI · updated 36m ago

Multimodal

Models that combine modalities such as vision-language models, video-language and any-to-any models.

537papers this week−27.2% vs last week

All papers · 3.3k With code · 219

Newest Most relevant

3.1k papers this month