Diffusion Transformer (DiT), a promising diffusion model for visual generation, demonstrates impressive performance but incurs significant computational overhead. Intriguingly, analysis of pre-trained DiT models reveals that global self-attention is often redundant, predominantly capturing local patterns—highlighting the potential for more efficient alternatives. In this paper, we revisit convolution as an alternative building block for constructing efficient and expressive diffusion models. However, naively replacing self-attention with convolution typically results in degraded performance. Our investigations attribute this performance gap to the higher channel redundancy in ConvNets compared to Transformers. To resolve this, we introduce a compact channel attention mechanism that promotes the activation of more diverse channels, thereby enhancing feature diversity. This leads to Diffusion ConvNet (DiCo), a family of diffusion models built entirely from standard ConvNet modules, offering strong generative performance with significant efficiency gains. On class-conditional ImageNet generation benchmarks, DiCo-XL achieves an FID of 2.05 at 256$\times$256 resolution and 2.53 at 512$\times$512, with a **2.7$\times$** and **3.1$\times$** speedup over DiT-XL/2, respectively. Furthermore, experimental results on MS-COCO demonstrate that the purely convolutional DiCo exhibits strong potential for text-to-image generation.

Low-shot image classification, where training images are limited or inaccessible, has benefited from recent progress on pretrained vision-language (VL) models with strong generalizability. e.g. CLIP. Prompt learning methods built with VL models generate text features from the class names that only have confined class-specific information. Large Language Models (LLMs), with their vast en-cyclopedic knowledge, emerge as the complement. Thus, in this paper, we discuss the integration of LLMs to enhance pretrained VL models, specifically on low-shot classification. However, the domain gap between language and vision blocks the direct application of LLMs. Thus, we propose LLaMp, Large Language Models as Prompt learners, that produces adaptive prompts for the CLIP text encoder, establishing it as the connecting bridge. Experiments show that, compared with other state-of-the-art prompt learning methods, LLaMP yields better performance on both zero-shot generalization and few-shot image classification, over a spectrum of 11 datasets. Code will be made available at: https://github.com/zhaohengz/LLaMP.

Large-scale pre-trained vision-language models (VLM) such as CLIP [32] have demonstrated noteworthy zero-shot classification capability, achieving 76.3% top-1 accuracy on ImageNet without seeing any examples. However, while applying CLIP to a downstream target domain, the presence of visual and text domain gaps and cross-modality misalignment can greatly impact the model performance. To address such challenges, we propose ReCLIP, a novel source-free domain adaptation method for VLMs, which does not require any source data or target labeled data. ReCLIP first learns a projection space to mitigate the misaligned visual-text embeddings and learns pseudo labels. Then, it deploys cross-modality self-training with the pseudo labels to update visual and text encoders, refine labels and reduce domain gaps and misalignment iteratively. With extensive experiments, we show that ReCLIP outperforms all the baselines significantly and improves the average accuracy of CLIP from 69.83% to 74.94% on 22 image classification benchmarks.

We propose a new setting for detecting unseen objects called Zero-shot Annotation object Detection (ZAD). It expands the zero-shot object detection setting by allowing the novel objects to exist in the training images and restricts the additional information the detector uses to novel category names. Recently, to detect unseen objects, largescale vision-language models (e.g., CLIP) are leveraged by different methods. The distillation-based methods have good overall performance but suffer from a long training schedule caused by two factors. First, existing work creates distillation regions biased to the base categories, which limits the distillation of novel category information. Second, directly using the raw feature from CLIP for distillation neglects the domain gap between the training data of CLIP and the detection datasets, which makes it difficult to learn the mapping from the image region to the vision-language feature space. To solve these problems, we propose Efficient feature distillation for Zero-shot Annotation object Detection (EZAD). Firstly, EZAD adapts the CLIP’s feature space to the target detection domain by re-normalizing CLIP; Secondly, EZAD uses CLIP to generate distillation proposals with potential novel category names to avoid the distillation being overly biased toward the base categories. Finally, EZAD takes advantage of semantic meaning for regression to further improve the model performance. As a result, EZAD outperforms the previous distillation-based methods in COCO by 4% with a much shorter training schedule and achieves a 3% improvement on the LVIS dataset. Our code is available at https://github.com/dragonlzm/EZAD

Existing work in physical robot caregiving is limited in its ability to provide long-term assistance. This is majorly due to (i) lack of well-defined problems, (ii) diversity of tasks, and (iii) limited access to stakeholders from the caregiving community. We propose Structuring Physically Assistive Robotics for Caregiving with Stakeholders-in-the-loop (SPARCS) to address these challenges. SPARCS is a framework for physical robot caregiving comprising (i) Building Blocks, models that define physical robot caregiving scenarios, (ii) Structured Workflows, hierarchical workflows that enable us to answer the Whats and Hows of physical robot caregiving, and (iii) SPARCS-box, a web-based platform to facilitate dialogue between all stakeholders. We collect clinical data for six care recipients with varying disabilities and demonstrate the use of SPARCS in designing well-defined caregiving scenarios and identifying their care requirements. All the data and workflows are available on SPARCS-box. We demonstrate the utility of SPARCS in building a robot-assisted feeding system for one of the care recipients. We also perform experiments to show the adaptability of this system to different caregiving scenarios. Finally, we identify open challenges in physical robot caregiving by consulting care recipients and caregivers. Supplementary material can be found at emprise.cs.cornell.edu/sparcs.

A common classification task situation is where one has a large amount of data available for training, but only a small portion is annotated with class labels. The goal of semi-supervised training, in this context, is to improve classification accuracy by leverage information not only from labeled data but also from a large amount of unlabeled data. Recent works have developed significant improvements by exploring the consistency constrain between differently augmented labeled and unlabeled data. Following this path, we propose a novel unsupervised objective that focuses on the less studied relationship between the high confidence unlabeled data that are similar to each other. The new proposed Pair Loss minimizes the statistical distance between high confidence pseudo labels with similarity above a certain threshold. Combining the Pair Loss with the techniques developed by the MixMatch family, our proposed SimPLE algorithm shows significant performance gains over previous algorithms on CIFAR-100 and Mini-ImageNet, and is on par with the state-of-the-art methods on CIFAR-10 and SVHN. Furthermore, SimPLE also outperforms the state-of-the-art methods in the transfer learning setting, where models are initialized by the weights pre-trained on ImageNet or DomainNet-Real. The code is available at github.com/zijian-hu/SimPLE.

We propose an unsupervised hashing method, exploiting a shallow neural network, that aims to produce binary codes that preserve the ranking induced by an original real-valued representation. This is motivated by the emergence of small-world graph-based approximate search methods that rely on local neighborhood ranking. We formalize the training process in an intuitive way by considering each training sample as a query and aiming to obtain a ranking of a random subset of the training set using the hash codes that is the same as the ranking using the original features. We also explore the use of a decoder to obtain an approximated reconstruction of the original features. At test time, we retrieve the most promising database samples using only the hash codes and perform re-ranking using the reconstructed features, thus allowing the complete elimination of the original real-valued features and the associated high memory cost. Experiments conducted on publicly available large-scale datasets show that our method consistently outperforms all compared state-of-the-art unsupervised hashing methods and that the reconstruction procedure can effectively boost the search accuracy with a minimal constant additional cost.

This paper presents a robust and efficient method for license plate detection with the purpose of accurately localizing vehicle license plates from complex scenes in real time. A simple yet effective image downscaling method is first proposed to substantially accelerate license plate localization without sacrificing detection performance compared with that achieved using the original image. Furthermore, a novel line density filter approach is proposed to extract candidate regions, thereby significantly reducing the area to be analyzed for license plate localization. Moreover, a cascaded license plate classifier based on linear support vector machines using color saliency features is introduced to identify the true license plate from among the candidate regions. For performance evaluation, a data set consisting of 3977 images captured from diverse scenes under different conditions is also presented. Extensive experiments on the widely used Caltech license plate data set and our newly introduced data set demonstrate that the proposed approach substantially outperforms state-of-the-art methods in terms of both detection accuracy and run-time efficiency, increasing the detection ratio from 91.09% to 96.62% while decreasing the run time from 672 to 42 ms for processing an image with a resolution of 1082×728 . The executable code and our collected data set are publicly available.