Especially in the ImageNet-1k classification, DTEM achieves a 37.2% reduction in

Recent token reduction methods for Vision Transformers (ViTs) incorporate token merging, which measures the similarities between token embeddings and combines the most similar pairs.However, their merging policies are directly dependent on intermediate features in ViTs, which prevents exploiting features tailored for merging and requires end-to-end training to improve token merging.In this paper, we propose Decoupled Token Embedding for Merging (DTEM) that enhances token merging through a decoupled embedding learned via a continuously relaxed token merging process.Our method introduces a lightweight embedding module decoupled from the ViT forward pass to extract dedicated features for token merging, thereby addressing the restriction from using intermediate features.The continuously relaxed token merging, applied during training, enables us to learn the decoupled embeddings in a differentiable manner.Thanks to the decoupled structure, our method can be seamlessly integrated into existing ViT backbones and trained either modularly by learning only the decoupled embeddings or end-to-end by fine-tuning. We demonstrate the applicability of DTEM on various tasks, including classification, captioning, and segmentation, with consistent improvement in token merging.Especially in the ImageNet-1k classification, DTEM achieves a 37.2\% reduction in FLOPs while maintaining a top-1 accuracy of 79.85\% with DeiT-small.

We propose a novel graph cross network (GXN) to achieve comprehensive feature learning from multiple scales of a graph. Based on trainable hierarchical representations of a graph, GXN enables the interchange of intermediate features across scales to promote information flow. Two key ingredients of GXN include a novel vertex infomax pooling (VIPool), which creates multiscale graphs in a trainable manner, and a novel feature-crossing layer, enabling feature interchange across scales. The proposed VIPool selects the most informative subset of vertices based on the neural estimation of mutual information between vertex features and neighborhood features. The intuition behind is that a vertex is informative when it can maximally reflect its neighboring information. The proposed feature-crossing layer fuses intermediate features between two scales for mutual enhancement by improving information flow and enriching multiscale features at hidden layers.

Recent Convolutional Neural Networks (CNNs) have achieved significant success by stacking multiple convolutional blocks, named procedures in this paper, to extract semantic features. However, they use the same procedure sequence for all inputs, regardless of the intermediate features.This paper proffers a simple yet effective idea of constructing parallel procedures and assigning similar intermediate features to the same specialized procedures in a divide-and-conquer fashion. It relieves each procedure's learning difficulty and thus leads to superior performance. Specifically, we propose a routing-by-memory mechanism for existing CNN architectures. In each stage of the network, we introduce parallel Procedural Units (PUs).

As Vision-Language Models (VLMs) are increasingly deployed in split-DNN configurations--with visual encoders (e.g., ResNet, ViT) operating on user devices and sending intermediate features to the cloud--there is a growing privacy risk from semantic information leakage. Existing approaches to reconstructing images from these intermediate features often result in blurry, semantically ambiguous images. To directly address semantic leakage, we propose CapRecover, a cross-modality inversion framework that recovers high-level semantic content, such as labels or captions, directly from intermediate features without image reconstruction. We evaluate CapRecover on multiple datasets and victim models, demonstrating strong performance in semantic recovery. Specifically, CapRecover achieves up to 92.71% Top-1 label accuracy on CIFAR-10 and generates fluent captions from ResNet50 features on COCO2017 with ROUGE-L scores up to 0.52. Our analysis further reveals that deeper convolutional layers encode significantly more semantic information compared to shallow layers. To mitigate semantic leakage, we introduce a simple yet effective protection method: adding random noise to intermediate features at each layer and removing the noise in the next layer. Experimental results show that this approach prevents semantic leakage without additional training costs. Our code is available at https://jus1mple.github.io/Image2CaptionAttack.

Speculative decoding (SD) has emerged as a powerful method for accelerating autoregressive generation in large language models (LLMs), yet its integration into vision-language models (VLMs) remains underexplored. We introduce DREAM, a novel speculative decoding framework tailored for VLMs that combines three key innovations: (1) a cross-attention-based mechanism to inject intermediate features from the target model into the draft model for improved alignment, (2) adaptive intermediate feature selection based on attention entropy to guide efficient draft model training, and (3) visual token compression to reduce draft model latency. DREAM enables efficient, accurate, and parallel multimodal decoding with significant throughput improvement. Experiments across a diverse set of recent popular VLMs, including LLaVA, Pixtral, SmolVLM and Gemma3, demonstrate up to 3.6x speedup over conventional decoding and significantly outperform prior SD baselines in both inference throughput and speculative draft acceptance length across a broad range of multimodal benchmarks. The code is publicly available at: https://github.com/SAI-Lab-NYU/DREAM.git