Osamu Yoshie

Quan Cui, Bingchen Zhao, Zhao-Min Chen et al.

This work simultaneously considers the discriminability and transferability properties of deep representations in the typical supervised learning task, i.e., image classification. By a comprehensive temporal analysis, we observe a trade-off between these two properties. The discriminability keeps increasing with the training progressing while the transferability intensely diminishes in the later training period. From the perspective of information-bottleneck theory, we reveal that the incompatibility between discriminability and transferability is attributed to the over-compression of input information. More importantly, we investigate why and how the InfoNCE loss can alleviate the over-compression, and further present a learning framework, named contrastive temporal coding~(CTC), to counteract the over-compression and alleviate the incompatibility. Extensive experiments validate that CTC successfully mitigates the incompatibility, yielding discriminative and transferable representations. Noticeable improvements are achieved on the image classification task and challenging transfer learning tasks. We hope that this work will raise the significance of the transferability property in the conventional supervised learning setting. Code is available at https://github.com/DTennant/dt-tradeoff.

Bingchen Zhao, Quan Cui, Hao Wu et al.

Many self-supervised learning methods are pre-trained on the well-curated ImageNet-1K dataset. In this work, given the excellent scalability of web data, we consider self-supervised pre-training on noisy web sourced image-text paired data. First, we conduct a benchmark study of representative self-supervised pre-training methods on large-scale web data in a like-for-like setting. We compare a range of methods, including single-modal ones that use masked training objectives and multi-modal ones that use image-text constrastive training. We observe that existing multi-modal methods do not outperform their single-modal counterparts on vision transfer learning tasks. We derive an information-theoretical view to explain these benchmark results, which provides insight into how to design a novel vision learner. Inspired by this insight, we present a new visual representation pre-training method, MUlti-modal Generator~(MUG), that learns from scalable web sourced image-text data. MUG achieves state-of-the-art transfer performance on a variety of tasks and demonstrates promising scaling properties. Pre-trained models and code will be made public upon acceptance.

Haolong Yan, Yeqing Shen, Xin Huang et al.

With the rapid development of Large Vision Language Models, the focus of Graphical User Interface (GUI) agent tasks shifts from single-screen tasks to complex screen navigation challenges. However, real-world GUI environments, such as PC software and mobile Apps, are often complex and proprietary, making it difficult to obtain the comprehensive environment information needed for agent training and evaluation. This limitation hinders systematic investigation and benchmarking of agent navigation capabilities. To address this limitation, we introduce GUI Exploration Lab, a simulation environment engine for GUI agent navigation research that enables flexible definition and composition of screens, icons, and navigation graphs, while providing full access to environment information for comprehensive agent training and evaluation. Through extensive experiments, we find that supervised fine-tuning enables effective memorization of fundamental knowledge, serving as a crucial foundation for subsequent training. Building on this, single-turn reinforcement learning further enhances generalization to unseen scenarios. Finally, multi-turn reinforcement learning encourages the development of exploration strategies through interactive trial and error, leading to further improvements in screen navigation performance. We validate our methods on both static and interactive benchmarks, demonstrating that our findings generalize effectively to real-world scenarios. These findings demonstrate the advantages of reinforcement learning approaches in GUI navigation and offer practical guidance for building more capable and generalizable GUI agents.

Quan Cui, Boyan Zhou, Yu Guo et al.

Pioneering dual-encoder pre-training works (e.g., CLIP and ALIGN) have revealed the potential of aligning multi-modal representations with contrastive learning. However, these works require a tremendous amount of data and computational resources (e.g., billion-level web data and hundreds of GPUs), which prevent researchers with limited resources from reproduction and further exploration. To this end, we propose a stack of novel methods, which significantly cut down the heavy resource dependency and allow us to conduct dual-encoder multi-modal representation alignment with limited resources. Besides, we provide a reproducible baseline of competitive results, namely ZeroVL, with only 14M publicly accessible academic datasets and 8 V100 GPUs. Additionally, we collect 100M web data for pre-training, and achieve comparable or superior results than state-of-the-art methods, further proving the effectiveness of our methods on large-scale data. We hope that this work will provide useful data points and experience for future research in contrastive vision-language pre-training. Code is available at https://github.com/zerovl/ZeroVL.

Mingshu Cai, Osamu Yoshie, Yuya Ieiri

Modern surveillance systems increasingly rely on multi-wavelength sensors and deep neural networks to recognize faces in infrared images captured at night. However, most facial recognition models are trained on visible light datasets, leading to substantial performance degradation on infrared inputs due to significant domain shifts. Early feature-based methods for infrared face recognition proved ineffective, prompting researchers to adopt generative approaches that convert infrared images into visible light images for improved recognition. This paradigm, known as Heterogeneous Face Recognition (HFR), faces challenges such as model and modality discrepancies, leading to distortion and feature loss in generated images. To address these limitations, this paper introduces a novel latent diffusion-based model designed to generate high-quality visible face images from thermal inputs while preserving critical identity features. A multi-attribute classifier is incorporated to extract key facial attributes from visible images, mitigating feature loss during infrared-to-visible image restoration. Additionally, we propose the Self-attn Mamba module, which enhances global modeling of cross-modal features and significantly improves inference speed. Experimental results on two benchmark datasets demonstrate the superiority of our approach, achieving state-of-the-art performance in both image quality and identity preservation.

Mingshu Cai, Yixuan Li, Osamu Yoshie et al.

Large-scale text-to-image diffusion models have achieved unprecedented success in image generation and editing. However, extending this success to video editing remains challenging. Recent video editing efforts have adapted pretrained text-to-image models by adding temporal attention mechanisms to handle video tasks. Unfortunately, these methods continue to suffer from temporal inconsistency issues and high computational overheads. In this study, we propose FluencyVE, which is a simple yet effective one-shot video editing approach. FluencyVE integrates the linear time-series module, Mamba, into a video editing model based on pretrained Stable Diffusion models, replacing the temporal attention layer. This enables global frame-level attention while reducing the computational costs. In addition, we employ low-rank approximation matrices to replace the query and key weight matrices in the causal attention, and use a weighted averaging technique during training to update the attention scores. This approach significantly preserves the generative power of the text-to-image model while effectively reducing the computational burden. Experiments and analyses demonstrate promising results in editing various attributes, subjects, and locations in real-world videos.

Cong Guan, Osamu Yoshie

Existing deraining models process all rainy images within a single network. However, different rain patterns have significant variations, which makes it challenging for a single network to handle diverse types of raindrops and streaks. To address this limitation, we propose a novel CLIP-driven rain perception network (CLIP-RPN) that leverages CLIP to automatically perceive rain patterns by computing visual-language matching scores and adaptively routing to sub-networks to handle different rain patterns, such as varying raindrop densities, streak orientations, and rainfall intensity. CLIP-RPN establishes semantic-aware rain pattern recognition through CLIP's cross-modal visual-language alignment capabilities, enabling automatic identification of precipitation characteristics across different rain scenarios. This rain pattern awareness drives an adaptive subnetwork routing mechanism where specialized processing branches are dynamically activated based on the detected rain type, significantly enhancing the model's capacity to handle diverse rainfall conditions. Furthermore, within sub-networks of CLIP-RPN, we introduce a mask-guided cross-attention mechanism (MGCA) that predicts precise rain masks at multi-scale to facilitate contextual interactions between rainy regions and clean background areas by cross-attention. We also introduces a dynamic loss scheduling mechanism (DLS) to adaptively adjust the gradients for the optimization process of CLIP-RPN. Compared with the commonly used $l_1$ or $l_2$ loss, DLS is more compatible with the inherent dynamics of the network training process, thus achieving enhanced outcomes. Our method achieves state-of-the-art performance across multiple datasets, particularly excelling in complex mixed datasets.