To Preserve or To Compress: An In-Depth Study of Connector Selection in Multimodal Large Language Models

In recent years, multimodal large language models (MLLMs) have garnered significant attention from both industry and academia. However, there is still considerable debate on constructing MLLM architectures, particularly regarding the selection of appropriate connectors for perception tasks of varying granularities. This paper systematically investigates the impact of connectors on MLLM performance. Specifically, we classify connectors into feature-preserving and feature-compressing types. Utilizing a unified classification standard, we categorize sub-tasks from three comprehensive benchmarks, MMBench, MME, and SEED-Bench, into three task types: coarse-grained perception, fine-grained perception, and reasoning, and evaluate the performance. Our findings reveal that feature-preserving connectors excel in \emph{fine-grained perception} tasks due to their ability to retain detailed visual information. In contrast, feature-compressing connectors, while less effective in fine-grained perception tasks, offer significant speed advantages and perform comparably in \emph{coarse-grained perception} and \emph{reasoning} tasks. These insights are crucial for guiding MLLM architecture design and advancing the optimization of MLLM architectures.

Recurrent Context Compression: Efficiently Expanding the Context Window of LLM

To extend the context length of Transformer-based large language models (LLMs) and improve comprehension capabilities, we often face limitations due to computational resources and bounded memory storage capacity. This work introduces a method called Recurrent Context Compression (RCC), designed to efficiently expand the context window length of LLMs within constrained storage space. We also investigate the issue of poor model responses when both instructions and context are compressed in downstream tasks, and propose an instruction reconstruction method to mitigate this problem. We validated the effectiveness of our approach on multiple tasks, achieving a compression rate of up to 32x on text reconstruction tasks with a BLEU4 score close to 0.95, and nearly 100\% accuracy on a passkey retrieval task with a sequence length of 1M. Finally, our method demonstrated competitive performance in long-text question-answering tasks compared to non-compressed methods, while significantly saving storage resources in long-text inference tasks. Our code, models, and demo are available at https://github.com/WUHU-G/RCC_Transformer

Adaptive Rentention & Correction for Continual Learning

Continual learning, also known as lifelong learning or incremental learning, refers to the process by which a model learns from a stream of incoming data over time. A common problem in continual learning is the classification layer's bias towards the most recent task. Traditionally, methods have relied on incorporating data from past tasks during training to mitigate this issue. However, the recent shift in continual learning to memory-free environments has rendered these approaches infeasible. In this study, we propose a solution focused on the testing phase. We first introduce a simple Out-of-Task Detection method, OTD, designed to accurately identify samples from past tasks during testing. Leveraging OTD, we then propose: (1) an Adaptive Retention mechanism for dynamically tuning the classifier layer on past task data; (2) an Adaptive Correction mechanism for revising predictions when the model classifies data from previous tasks into classes from the current task. We name our approach Adaptive Retention & Correction (ARC). While designed for memory-free environments, ARC also proves effective in memory-based settings. Extensive experiments show that our proposed method can be plugged in to virtually any existing continual learning approach without requiring any modifications to its training procedure. Specifically, when integrated with state-of-the-art approaches, ARC achieves an average performance increase of 2.7% and 2.6% on the CIFAR-100 and Imagenet-R datasets, respectively.

A Survey on Video Diffusion Models

The recent wave of AI-generated content (AIGC) has witnessed substantial success in computer vision, with the diffusion model playing a crucial role in this achievement. Due to their impressive generative capabilities, diffusion models are gradually superseding methods based on GANs and auto-regressive Transformers, demonstrating exceptional performance not only in image generation and editing, but also in the realm of video-related research. However, existing surveys mainly focus on diffusion models in the context of image generation, with few up-to-date reviews on their application in the video domain. To address this gap, this paper presents a comprehensive review of video diffusion models in the AIGC era. Specifically, we begin with a concise introduction to the fundamentals and evolution of diffusion models. Subsequently, we present an overview of research on diffusion models in the video domain, categorizing the work into three key areas: video generation, video editing, and other video understanding tasks. We conduct a thorough review of the literature in these three key areas, including further categorization and practical contributions in the field. Finally, we discuss the challenges faced by research in this domain and outline potential future developmental trends. A comprehensive list of video diffusion models studied in this survey is available at https://github.com/ChenHsing/Awesome-Video-Diffusion-Models.

Panoptic Vision-Language Feature Fields

Recently, methods have been proposed for 3D open-vocabulary semantic segmentation. Such methods are able to segment scenes into arbitrary classes given at run-time using their text description. In this paper, we propose to our knowledge the first algorithm for open-vocabulary panoptic segmentation, simultaneously performing both semantic and instance segmentation. Our algorithm, Panoptic Vision-Language Feature Fields (PVLFF) learns a feature field of the scene, jointly learning vision-language features and hierarchical instance features through a contrastive loss function from 2D instance segment proposals on input frames. Our method achieves comparable performance against the state-of-the-art close-set 3D panoptic systems on the HyperSim, ScanNet and Replica dataset and outperforms current 3D open-vocabulary systems in terms of semantic segmentation. We additionally ablate our method to demonstrate the effectiveness of our model architecture. Our code will be available at https://github.com/ethz-asl/autolabel.

SPPNet: A Single-Point Prompt Network for Nuclei Image Segmentation

Image segmentation plays an essential role in nuclei image analysis. Recently, the segment anything model has made a significant breakthrough in such tasks. However, the current model exists two major issues for cell segmentation: (1) the image encoder of the segment anything model involves a large number of parameters. Retraining or even fine-tuning the model still requires expensive computational resources. (2) in point prompt mode, points are sampled from the center of the ground truth and more than one set of points is expected to achieve reliable performance, which is not efficient for practical applications. In this paper, a single-point prompt network is proposed for nuclei image segmentation, called SPPNet. We replace the original image encoder with a lightweight vision transformer. Also, an effective convolutional block is added in parallel to extract the low-level semantic information from the image and compensate for the performance degradation due to the small image encoder. We propose a new point-sampling method based on the Gaussian kernel. The proposed model is evaluated on the MoNuSeg-2018 dataset. The result demonstrated that SPPNet outperforms existing U-shape architectures and shows faster convergence in training. Compared to the segment anything model, SPPNet shows roughly 20 times faster inference, with 1/70 parameters and computational cost. Particularly, only one set of points is required in both the training and inference phases, which is more reasonable for clinical applications. The code for our work and more technical details can be found at https://github.com/xq141839/SPPNet.

PromptFusion: Decoupling Stability and Plasticity for Continual Learning

Continual learning refers to the capability of continuously learning from a stream of data. Current research mainly focuses on relieving catastrophic forgetting, and most of their success is at the cost of limiting the performance of newly incoming tasks. Such a trade-off is referred to as the stabilityplasticity dilemma and is a more general and challenging problem for continual learning. However, the inherent conflict between these two concepts makes it seemingly impossible to devise a satisfactory solution to both of them simultaneously. Therefore, we ask, "is it possible to divide them into two problems to conquer independently?" To this end, we propose a prompt-tuning-based method termed PromptFusion to enable the decoupling of stability and plasticity. Specifically, PromptFusion consists of a carefully designed Stabilizer module that deals with catastrophic forgetting and a Booster module to learn new knowledge concurrently. During training, PromptFusion first passes an input image to the two modules separately. Then the resulting logits are further fused with a learnable weight parameter. Finally, a weight mask is applied to the derived logits to balance between old and new classes. Extensive experiments show that our method achieves promising results on popular continual learning datasets for both class-incremental and domain incremental settings. Especially on Split-Imagenet-R, one of the most challenging datasets for class-incremental learning, our method exceeds state-of-the-art prompt-based methods L2P and DualPrompt by more than 10%.

Multi-Prompt Alignment for Multi-source Unsupervised Domain Adaptation

Most existing methods for multi-source unsupervised domain adaptation (UDA) rely on a common feature encoder to extract domain-invariant features. However, learning such an encoder involves updating the parameters of the entire network, which makes the optimization computationally expensive, particularly when coupled with min-max objectives. Inspired by recent advances in prompt learning that adapts high-capacity deep models for downstream tasks in a computationally economic way, we introduce Multi-Prompt Alignment (MPA), a simple yet efficient two-stage framework for multi-source UDA. Given a source and target domain pair, MPA first trains an individual prompt to minimize the domain gap through a contrastive loss, while tuning only a small set of parameters. Then, MPA derives a low-dimensional latent space through an auto-encoding process that maximizes the agreement of multiple learned prompts. The resulting embedding further facilitates generalization to unseen domains. Extensive experiments show that our method achieves state-of-the-art results on popular benchmark datasets while requiring substantially fewer tunable parameters. To the best of our knowledge, we are the first to apply prompt learning to the multi-source UDA problem and our method achieves the highest reported average accuracy of 54.1% on DomainNet, the most challenging UDA dataset to date, with only 15.9M parameters trained. More importantly, we demonstrate that the learned embedding space can be easily adapted to novel unseen domains.

Facial Attribute Transformers for Precise and Robust Makeup Transfer

In this paper, we address the problem of makeup transfer, which aims at transplanting the makeup from the reference face to the source face while preserving the identity of the source. Existing makeup transfer methods have made notable progress in generating realistic makeup faces, but do not perform well in terms of color fidelity and spatial transformation. To tackle these issues, we propose a novel Facial Attribute Transformer (FAT) and its variant Spatial FAT for high-quality makeup transfer. Drawing inspirations from the Transformer in NLP, FAT is able to model the semantic correspondences and interactions between the source face and reference face, and then precisely estimate and transfer the facial attributes. To further facilitate shape deformation and transformation of facial parts, we also integrate thin plate splines (TPS) into FAT, thus creating Spatial FAT, which is the first method that can transfer geometric attributes in addition to color and texture. Extensive qualitative and quantitative experiments demonstrate the effectiveness and superiority of our proposed FATs in the following aspects: (1) ensuring high-fidelity color transfer; (2) allowing for geometric transformation of facial parts; (3) handling facial variations (such as poses and shadows) and (4) supporting high-resolution face generation.

Annotation Cleaning for the MSR-Video to Text Dataset

The video captioning task is to describe the video contents with natural language by the machine. Many methods have been proposed for solving this task. A large dataset called MSR Video to Text (MSR-VTT) is often used as the benckmark dataset for testing the performance of the methods. However, we found that the human annotations, i.e., the descriptions of video contents in the dataset are quite noisy, e.g., there are many duplicate captions and many captions contain grammatical problems. These problems may pose difficulties to video captioning models for learning. We cleaned the MSR-VTT annotations by removing these problems, then tested several typical video captioning models on the cleaned dataset. Experimental results showed that data cleaning boosted the performances of the models measured by popular quantitative metrics. We recruited subjects to evaluate the results of a model trained on the original and cleaned datasets. The human behavior experiment demonstrated that trained on the cleaned dataset, the model generated captions that were more coherent and more relevant to contents of the video clips. The cleaned dataset is publicly available.