NeurIPS 2025 Papers

While vision language models (VLMs) excel in 2D semantic visual understanding, their ability to quantitatively reason about 3D spatial relationships remains underexplored due to the deficiency of spatial representation ability of 2D images. In this paper, we analyze the problem hindering VLMs’ spatial understanding abilities and propose SD-VLM, a novel framework that significantly enhances fundamental spatial perception abilities of VLMs through two key contributions: (1) propose Massive Spatial Measuring and Understanding (MSMU) dataset with precise spatial annotations, and (2) introduce a simple depth positional encoding method strengthening VLMs’ spatial awareness. MSMU dataset includes massive quantitative spatial tasks with 700K QA pairs, 2.5M physical numerical annotations, and 10K chain-of-thought augmented samples. We have trained SD-VLM, a strong generalist VLM which shows superior quantitative spatial measuring and understanding capability. SD-VLM not only achieves state-of-the-art performance on our proposed MSMU-Bench, but also shows spatial generalization abilities on other spatial understanding benchmarks including Q-Spatial and SpatialRGPTBench. Extensive experiments demonstrate that SD-VLM outperforms GPT-4o and Intern-VL3-78B by 26.91% and 25.56% respectively on MSMU-Bench. Code and models are released at https://github.com/cpystan/SD-VLM.

Conventional Vision Transformer streamlines visual modeling by employing a uniform input resolution, which underestimates the inherent variability of natural visual data and incurs a cost in spatial-contextual fidelity. While preliminary explorations have superficially investigated native resolution modeling, existing works still lack systematic training recipe from the visual representation perspective. To bridge this gap, we introduce Unified Vision Transformer with Native Resolution, i.e. UniViTAR, a family of homogeneous vision foundation models tailored for unified visual modality and native resolution scenario in the era of multimodal. Our framework first conducts architectural upgrades to the vanilla paradigm by integrating multiple advanced components. Building upon these improvements, a progressive training paradigm is introduced, which strategically combines two core mechanisms: (1) resolution curriculum learning, transitioning from fixed-resolution pretraining to native resolution tuning, thereby leveraging ViT’s inherent adaptability to variable-length sequences, and (2) visual modality adaptation via inter-batch image-video switching, which balances computational efficiency with enhanced temporal reasoning. In parallel, a hybrid training framework further synergizes sigmoid-based contrastive loss with feature distillation from a frozen teacher model, thereby accelerating early-stage convergence. Finally, trained exclusively on public accessible image-caption data, our UniViTAR family across multiple model scales from 0.3B to 1B achieves state-of-the-art performance on a wide variety of visual-related tasks. The code and models are available here.

We present a Directional Consistency (DC)-driven Adaptive Density Control (ADC) for 3D Gaussian Splatting (DC4GS). Whereas the conventional ADC bases its primitive splitting on the magnitudes of positional gradients, we further incorporate the DC of the gradients into ADC, and realize it through the angular coherence of the gradients. Our DC better captures local structural complexities in ADC, avoiding redundant splitting. When splitting is required, we again utilize the DC to define optimal split positions so that sub-primitives best align with the local structures than the conventional random placement. As a consequence, our DC4GS greatly reduces the number of primitives (up to 30\% in our experiments) than the existing ADC, and also enhances reconstruction fidelity greatly.

Recent studies on Vision-Language-Action (VLA) models have shifted from the end-to-end action-generation paradigm toward a pipeline involving task planning followed by action generation, demonstrating improved performance on various complex, long-horizon manipulation tasks. However, existing approaches vary significantly in terms of network architectures, planning paradigms, representations, and training data sources, making it challenging for researchers to identify the precise sources of performance gains and determine which component is more difficult to learn. To systematically investigate the impacts of different planning paradigms and representations isolating from network architectures and training data, in this paper, we introduce \name, a unified VLA architecture suite capable of various task planning paradigms, and design a comprehensive suite of controlled experiments across diverse object categories (rigid and deformable), visual modalities (2D and 3D), environments (simulation and real-world), and end-effectors (grippers and dexterous hands). Our results demonstrate that: 1) visually grounded planning representations are generally better than language planning representations; 2) the Hierarchical-VLA paradigm generally achieves superior performance than other paradigms, albeit at the cost of slower training and inference speeds.

With the rapid discovery of emergent phenomena in deep learning and large language models, understanding their cause has become an urgent need. Here, we propose a rigorous entropic-force theory for understanding the learning dynamics of neural networks trained with stochastic gradient descent (SGD) and its variants. Building on the theory of parameter symmetries and an entropic loss landscape, we show that representation learning is crucially governed by emergent entropic forces arising from stochasticity and discrete-time updates. These forces systematically break continuous parameter symmetries and preserve discrete ones, leading to a series of gradient balance phenomena that resemble the equipartition property of thermal systems. These phenomena, in turn, (a) explain the universal alignment of neural representations between AI models and lead to a proof of the Platonic Representation Hypothesis, and (b) reconcile the seemingly contradictory observations of sharpness- and flatness-seeking behavior of deep learning optimization. Our theory and experiments demonstrate that a combination of entropic forces and symmetry breaking is key to understanding emergent phenomena in deep learning.

We introduce a comprehensive framework for modeling single cell transcriptomic responses to perturbations, aimed at standardizing benchmarking in this rapidly evolving field. Our approach includes a modular and user-friendly model development and evaluation platform, a collection of diverse perturbational datasets, and a set of metrics designed to fairly compare models and dissect their performance. Through extensive evaluation of both published and baseline models across diverse datasets, we highlight the limitations of widely used models, such as mode collapse. We also demonstrate the importance of rank metrics which complement traditional model fit measures, such as RMSE, for validating model effectiveness. Notably, our results show that while no single model architecture clearly outperforms others, simpler architectures are generally competitive and scale well with larger datasets. Overall, this benchmarking exercise sets new standards for model evaluation, supports robust model development, and furthers the use of these models to simulate genetic and chemical screens for therapeutic discovery.

Pre-training has proven effective in addressing data scarcity and performance limitations in solving PDE problems with neural operators. However, challenges remain due to the heterogeneity of PDE datasets in equation types, which leads to high errors in mixed training. Additionally, dense pre-training models that scale parameters by increasing network width or depth incur significant inference costs. To tackle these challenges, we propose a novel **M**ixture-**o**f-**E**xperts **P**re-training **O**perator **T**ransformer (**MoE-POT**), a sparse-activated architecture that scales parameters efficiently while controlling inference costs. Specifically, our model adopts a layer-wise router-gating network to dynamically select 4 routed experts from 16 expert networks during inference, enabling the model to focus on equation-specific features. Meanwhile, we also integrate 2 shared experts, aiming to capture common properties of PDE and reduce redundancy among routed experts. The final output is computed as the weighted average of the results from all activated experts. We pre-train models with parameters from 30M to 0.5B on 6 public PDE datasets. Our model with 90M activated parameters achieves up to a 40\% reduction in zero-shot error compared with existing models with 120M activated parameters. Additionally, we conduct interpretability analysis, showing that dataset types can be inferred from router-gating network decisions, which validates the rationality and effectiveness of the MoE architecture.

Early-Exit Deep Neural Networks enable adaptive inference by allowing prediction at intermediary layers, significantly reducing computational costs and latency. Most of the early exit strategies greedily exit a sample at an intermediary layer if the confidence in class prediction exceeds a predefined threshold that is set using a static validation set. This is problematic as the model might be overconfident in a wrong class. Also, they are not robust to distribution shifts encountered in deployment, which can undermine model trustworthiness and accuracy. To address these challenges, we propose UAT that adapts the threshold for exit decisions using a Multi-Armed Bandit framework, enabling online, unsupervised adjustment of exit decisions. UAT makes decisions based on a new reward function that assesses predictive certainty and its reliability to balance computational efficiency and prediction quality while penalizing unnecessary late exits. We provide guarantees on risk achieved by UAT and validate its performance on diverse tasks spanning vision-language understanding, text generation, and classification. Our framework demonstrates consistent improvements in speedup $(1.70-2.10\times)$ with a minimal performance drop $(<2)$\% as compared to full model performance.

Time series, typically represented as numerical sequences, can also be transformed into images and texts, offering multi-modal views (MMVs) of the same underlying signal. These MMVs can reveal complementary patterns and enable the use of powerful pre-trained large models, such as large vision models (LVMs), for long-term time series forecasting (LTSF). However, as we identified in this work, the state-of-the-art (SOTA) LVM-based forecaster poses an inductive bias towards "forecasting periods". To harness this bias, we propose DMMV, a novel decomposition-based multi-modal view framework that leverages trend-seasonal decomposition and a novel backcast-residual based adaptive decomposition to integrate MMVs for LTSF. Comparative evaluations against 14 SOTA models across diverse datasets show that DMMV outperforms single-view and existing multi-modal baselines, achieving the best mean squared error (MSE) on 6 out of 8 benchmark datasets. The code for this paper is available at: https://github.com/D2I-Group/dmmv.

The optimal bit-width for achieving the best trade-off between quantized model size and accuracy has been a subject of ongoing debate. While some advocate for 4-bit quantization, others propose that 1.58-bit offers superior results. However, the lack of a cohesive framework for different bits has left such conclusions relatively tenuous. We present ParetoQ, the first unified framework that facilitates rigorous comparisons across 1-bit, 1.58-bit, 2-bit, 3-bit, and 4-bit quantization settings. Our findings reveal a notable learning transition between 2 and 3 bits: For 3-bits and above, the fine-tuned models stay close to their original pre-trained distributions, whereas for learning 2-bit networks or below, the representations change drastically. By optimizing training schemes and refining quantization functions, ParetoQ surpasses all previous methods tailored to specific bit widths. Remarkably, our ParetoQ ternary 600M-parameter model even outperforms the previous SoTA ternary 3B-parameter model in accuracy, using only one-fifth of the parameters. Extensive experimentation shows that ternary, 2-bit, and 3-bit quantization maintains comparable performance in the size-accuracy trade-off and generally exceeds 4-bit and binary quantization. Considering hardware constraints, 2-bit quantization offers promising potential for memory reduction and speedup.

Achieving better alignment between vision embeddings and Large Language Models (LLMs) is crucial for enhancing the abilities of Multimodal LLMs (MLLMs), particularly for recent models that rely on powerful pretrained vision encoders and LLMs. A common approach to connect the pretrained vision encoder and LLM is through a projector applied after the vision encoder. However, the projector is often trained to enable the LLM to generate captions, and hence the mechanism by which LLMs understand each vision token remains unclear. In this work, we first investigate the role of the projector in compressing vision embeddings and aligning them with word embeddings. We show that the projector significantly compresses visual information, removing redundant details while preserving essential elements necessary for the LLM to understand visual content. We then examine patch-level alignment---the alignment between each vision patch and its corresponding semantic words---and propose a $\textit{multi-semantic alignment hypothesis}$. Our analysis indicates that the projector trained by caption loss improves patch-level alignment but only to a limited extent, resulting in weak and coarse alignment. To address this issue, we propose $\textit{patch-aligned training}$ to efficiently enhance patch-level alignment. Our experiments show that patch-aligned training (1) achieves stronger compression capability and improved patch-level alignment, enabling the MLLM to generate higher-quality captions, (2) improves the MLLM's performance by 16% on referring expression grounding tasks, 4% on question-answering tasks, and 3% on modern instruction-following benchmarks when using the same supervised fine-tuning (SFT) setting. The proposed method can be easily extended to other multimodal models.

Recent advances in vision-language-action (VLA) models have shown promise in integrating image generation with action prediction to improve generalization and reasoning in robot manipulation. However, existing methods are limited to challenging image-based forecasting, which suffers from redundant information and lacks comprehensive and critical world knowledge, including geometry, semantics and spatial information. To address these limitations, we propose DreamVLA, a novel VLA framework that integrates comprehensive world knowledge forecasting to enable inverse dynamics modeling, thereby establishing an action-forecasting loop for manipulation tasks. Specifically, DreamVLA introduces a dynamic-region-guided world knowledge prediction mechanism, which anticipates visual, depth, geometric, semantic, and segmentation cues to provide compact yet comprehensive representations for action planning. This design aligns with how humans interact with the world by first forming abstract multimodal reasoning chains before acting. Moreover, to model the conditional distribution over future actions, we employ a diffusion-based transformer that disentangles action representations from shared latent features and better captures multimodal uncertainty. Extensive experiments on both real-world and simulation environments demonstrate that DreamVLA achieves 76.7% success rate on real robot tasks and 4.45 average length on the CALVIN ABC-D benchmarks.

Many machine learning tasks can benefit from external knowledge. Large knowledge graphs store such knowledge, and embedding methods can be used to distill it into ready-to-use vector representations for downstream applications. For this purpose, current models have however two limitations: they are primarily optimized for link prediction, via local contrastive learning, and their application to the largest graphs requires significant engineering effort due to GPU memory limits. To address these, we introduce SEPAL: a Scalable Embedding Propagation ALgorithm for large knowledge graphs designed to produce high-quality embeddings for downstream tasks at scale. The key idea of SEPAL is to ensure global embedding consistency by optimizing embeddings only on a small core of entities, and then propagating them to the rest of the graph with message passing. We evaluate SEPAL on 7 large-scale knowledge graphs and 46 downstream machine learning tasks. Our results show that SEPAL significantly outperforms previous methods on downstream tasks. In addition, SEPAL scales up its base embedding model, enabling fitting huge knowledge graphs on commodity hardware. Our code is available at: <https://github.com/flefebv/sepal.git>.

World models allow agents to simulate the consequences of actions in imagined environments for planning, control, and long-horizon decision-making. However, existing autoregressive world models struggle with visually coherent predictions due to disrupted spatial structure, inefficient decoding, and inadequate motion modeling. In response, we propose Scale-wise Autoregression with Motion PrOmpt (SAMPO), a hybrid framework that combines visual autoregressive modeling for intra-frame generation with causal modeling for next-frame generation. Specifically, SAMPO integrates temporal causal decoding with bidirectional spatial attention, which preserves spatial locality and supports parallel decoding within each scale. This design significantly enhances both temporal consistency and rollout efficiency. To further improve dynamic scene understanding, we devise an asymmetric multi-scale tokenizer that preserves spatial details in observed frames and extracts compact dynamic representations for future frames, optimizing both memory usage and model performance. Additionally, we introduce a trajectory-aware motion prompt module that injects spatiotemporal cues about object and robot trajectories, focusing attention on dynamic regions and improving temporal consistency and physical realism. Extensive experiments show that SAMPO achieves competitive performance in action-conditioned video prediction and model-based control, improving generation quality with 4.4× faster inference. We also evaluate SAMPO's zero-shot generalization and scaling behavior, demonstrating its ability to generalize to unseen tasks and benefit from larger model sizes.

Deep neural networks, particularly neural operators, provide an efficient alternative to costly simulations in surrogate modeling. However, their performance is often constrained by the need for large-scale labeled datasets, which are costly and challenging to acquire in many scientific domains. Semi-supervised learning reduces label reliance by leveraging unlabeled data yet remains vulnerable to noisy pseudo-labels that mislead training and undermine robustness. To address these challenges, we propose a novel framework, Uncertainty-Informed Meta Pseudo Labeling (UMPL). The core mechenism is to refine pseudo-label quality through uncertainty-informed feedback signals. Specifically, the teacher model generates pseudo labels via epistemic uncertainty, while the student model learns from these labels and provides feedback based on aleatoric uncertainty. This interplay forms a meta-learning loop where enhanced generalization and improved pseudo-label quality reinforce each other, enabling the student model to achieve more stable uncertainty estimation and leading to more robust training. Notably, This framework is model-agnostic and can be seamlessly integrated into various neural architectures, facilitating effective exploitation of unlabeled data to enhance generalization in distribution shifts and out-of-distribution scenarios. Extensive evaluations of four models across seven tasks covering steady state and transient prediction problems demonstrate that UMPL consistently outperforms the best existing semi-supervised regression methods. When using only 10% of the fully supervised training data, UMPL achieves a 14.18% improvement, highlighting its strong effectiveness under limited supervision. Our codes are available at https://github.com/small-dumpling/UMPL.

In this work, we present a novel direction to build an image tokenizer directly on top of a frozen vision foundation model, which is a largely underexplored area. Specifically, we employ a frozen vision foundation model as the encoder of our tokenizer. To enhance its effectiveness, we introduce two key components: (1) a region-adaptive quantization framework that reduces redundancy in the pre-trained features on regular 2D grids, and (2) a semantic reconstruction objective that aligns the tokenizer’s outputs with the foundation model’s representations to preserve semantic fidelity. Based on these designs, our proposed image tokenizer, \textbf{\ours}, achieves substantial improvements in image reconstruction and generation quality, while also enhancing token efficiency. It further boosts autoregressive (AR) generation---achieving a gFID of \textbf{1.36} on ImageNet benchmarks, while accelerating model convergence by \textbf{three times}, and enabling high-fidelity class-conditional synthesis without the need for classifier-free guidance (CFG). The code is available at \href{https://github.com/CVMI-Lab/VFMTok}{https://github.com/CVMI-Lab/VFMTok}.

Text-to-video diffusion models are notoriously limited in their ability to model temporal aspects such as motion, physics, and dynamic interactions. Existing approaches address this limitation by retraining the model or introducing external conditioning signals to enforce temporal consistency. In this work, we explore whether a meaningful temporal representation can be extracted directly from the predictions of a pre-trained model without any additional training or auxiliary inputs. We introduce __FlowMo__, a novel training-free guidance method that enhances motion coherence using only the model's own predictions in each diffusion step. FlowMo first derives an appearance-debiased temporal representation by measuring the distance between latents corresponding to consecutive frames. This highlights the implicit temporal structure predicted by the model. It then estimates motion coherence by measuring the patch-wise variance across the temporal dimension, and guides the model to reduce this variance dynamically during sampling. Extensive experiments across multiple text-to-video models demonstrate that FlowMo significantly improves motion coherence without sacrificing visual quality or prompt alignment, offering an effective plug-and-play solution for enhancing the temporal fidelity of pre-trained video diffusion models.

Multimodal Sentiment Analysis (MSA) aims to predict sentiment from diverse data types, such as video, audio, and language. Recent progress in Multimodal Large Language Models (MLLMs) have demonstrated impressive performance across various tasks. However, in MSA, the increase in computational costs does not always correspond to a significant improvement in performance, raising concerns about the cost-effectiveness of applying MLLMs to MSA. This paper introduces the MLLM-Guided Multimodal Sentiment Learning Framework (MMSLF). It improves the performance of task-specific MSA models by leveraging the generalized knowledge of MLLMs through a teacher-student framework, rather than directly using MLLMs for sentiment prediction. First, the proposed teacher built upon a powerful MLLM (e.g., GPT-4o-mini), guides the student model to align multimodal representations through MLLM-generated context-aware prompts. Then, knowledge distillation enables the student to mimic the teacher’s predictions, thus allowing it to predict sentiment independently without relying on the context-aware prompts. Extensive experiments on the SIMS, MOSI, and MOSEI datasets demonstrate that our framework enables task-specific models to achieve state-of-the-art performance across most metrics. This also provides new insights into the application of general MLLMs for improving MSA.

The parameter size of modern large language models (LLMs) can be scaled up to the trillion-level via the sparsely-activated Mixture-of-Experts (MoE) technique to avoid excessive increase of the computational costs. To further improve training efficiency, pipelining computation and communication has become a promising solution for distributed MoE training. However, existing work primarily focuses on scheduling tasks within the MoE layer, such as expert computing and all-to-all (A2A) communication, while neglecting other key operations including multi-head attention (MHA) computing, gating, and all-reduce communication. In this paper, we propose FlowMoE, a scalable framework for scheduling multi-type task pipelines. First, FlowMoE constructs a unified pipeline to consistently scheduling MHA computing, gating, expert computing, and A2A communication. Second, FlowMoE introduces a tensor chunk-based priority scheduling mechanism to overlap the all-reduce communication with all computing tasks. We implement FlowMoE as an adaptive and generic framework atop PyTorch. Extensive experiments with 675 typical MoE layers and four real-world MoE models across two GPU clusters demonstrate that our proposed FlowMoE framework outperforms state-of-the-art MoE training frameworks, reducing training time by14%-57%, energy consumption by 10%-39%, and memory usage by 7%-32%. FlowMoE’s code is anonymously available at https://anonymous.4open.science/r/FlowMoE.