No Other Representation Component Is Needed: Diffusion Transformers Can Provide Representation Guidance by Themselves

Recent studies have demonstrated that learning a meaningful internal representation can both accelerate generative training and enhance generation quality of the diffusion transformers. However, existing approaches necessitate to either introduce an additional and complex representation training framework or rely on a large-scale, pre-trained representation foundation model to provide representation guidance during the original generative training process. In this study, we posit that the unique discriminative process inherent to diffusion transformers enables them to offer such guidance without requiring external representation components. We therefore propose Self-Representation A}lignment (SRA), a simple yet straightforward method that obtain representation guidance through a self-distillation manner. Specifically, SRA aligns the output latent representation of the diffusion transformer in earlier layer with higher noise to that in later layer with lower noise to progressively enhance the overall representation learning during only generative training process. Experimental results indicate that applying SRA to DiTs and SiTs yields consistent performance improvements. Moreover, SRA not only significantly outperforms approaches relying on auxiliary, complex representation training frameworks but also achieves performance comparable to methods that heavily dependent on powerful external representation priors.

Vision-Language Model Fine-Tuning via Simple Parameter-Efficient Modification

Recent advances in fine-tuning Vision-Language Models (VLMs) have witnessed the success of prompt tuning and adapter tuning, while the classic model fine-tuning on inherent parameters seems to be overlooked. It is believed that fine-tuning the parameters of VLMs with few-shot samples corrupts the pre-trained knowledge since fine-tuning the CLIP model even degrades performance. In this paper, we revisit this viewpoint, and propose a new perspective: fine-tuning the specific parameters instead of all will uncover the power of classic model fine-tuning on VLMs. Through our meticulous study, we propose ClipFit, a simple yet effective method to fine-tune CLIP without introducing any overhead of extra parameters. We demonstrate that by only fine-tuning the specific bias terms and normalization layers, ClipFit can improve the performance of zero-shot CLIP by 7.27\% average harmonic mean accuracy. Lastly, to understand how fine-tuning in CLIPFit affects the pre-trained models, we conducted extensive experimental analyses w.r.t. changes in internal parameters and representations. We found that low-level text bias layers and the first layer normalization layer change much more than other layers. The code is available at \url{https://github.com/minglllli/CLIPFit}.

Structure Gaussian SLAM with Manhattan World Hypothesis

Gaussian SLAM systems have made significant advancements in improving the efficiency and fidelity of real-time reconstructions. However, these systems often encounter incomplete reconstructions in complex indoor environments, characterized by substantial holes due to unobserved geometry caused by obstacles or limited view angles. To address this challenge, we present Manhattan Gaussian SLAM (MG-SLAM), an RGB-D system that leverages the Manhattan World hypothesis to enhance geometric accuracy and completeness. By seamlessly integrating fused line segments derived from structured scenes, MG-SLAM ensures robust tracking in textureless indoor areas. Moreover, The extracted lines and planar surface assumption allow strategic interpolation of new Gaussians in regions of missing geometry, enabling efficient scene completion. Extensive experiments conducted on both synthetic and real-world scenes demonstrate that these advancements enable our method to achieve state-of-the-art performance, marking a substantial improvement in the capabilities of Gaussian SLAM systems.

CEIR: Concept-based Explainable Image Representation Learning

In modern machine learning, the trend of harnessing self-supervised learning to derive high-quality representations without label dependency has garnered significant attention. However, the absence of label information, coupled with the inherently high-dimensional nature, improves the difficulty for the interpretation of learned representations. Consequently, indirect evaluations become the popular metric for evaluating the quality of these features, leading to a biased validation of the learned representation rationale. To address these challenges, we introduce a novel approach termed Concept-based Explainable Image Representation (CEIR). Initially, using the Concept-based Model (CBM) incorporated with pretrained CLIP and concepts generated by GPT-4, we project input images into a concept vector space. Subsequently, a Variational Autoencoder (VAE) learns the latent representation from these projected concepts, which serves as the final image representation. Due to the capability of the representation to encapsulate high-level, semantically relevant concepts, the model allows for attributions to a human-comprehensible concept space. This not only enhances interpretability but also preserves the robustness essential for downstream tasks. For instance, our method exhibits state-of-the-art unsupervised clustering performance on benchmarks such as CIFAR10, CIFAR100, and STL10. Furthermore, capitalizing on the universality of human conceptual understanding, CEIR can seamlessly extract the related concept from open-world images without fine-tuning. This offers a fresh approach to automatic label generation and label manipulation.