RemDet: Rethinking Efficient Model Design for UAV Object Detection

Object detection in Unmanned Aerial Vehicle (UAV) images has emerged as a focal area of research, which presents two significant challenges: i) objects are typically small and dense within vast images; ii) computational resource constraints render most models unsuitable for real-time deployment. Current real-time object detectors are not optimized for UAV images, and complex methods designed for small object detection often lack real-time capabilities. To address these challenges, we propose a novel detector, RemDet (Reparameter efficient multiplication Detector). Our contributions are as follows: 1) Rethinking the challenges of existing detectors for small and dense UAV images, and proposing information loss as a design guideline for efficient models. 2) We introduce the ChannelC2f module to enhance small object detection performance, demonstrating that high-dimensional representations can effectively mitigate information loss. 3) We design the GatedFFN module to provide not only strong performance but also low latency, effectively addressing the challenges of real-time detection. Our research reveals that GatedFFN, through the use of multiplication, is more cost-effective than feed-forward networks for high-dimensional representation. 4) We propose the CED module, which combines the advantages of ViT and CNN downsampling to effectively reduce information loss. It specifically enhances context information for small and dense objects. Extensive experiments on large UAV datasets, Visdrone and UAVDT, validate the real-time efficiency and superior performance of our methods. On the challenging UAV dataset VisDrone, our methods not only provided state-of-the-art results, improving detection by more than 3.4%, but also achieve 110 FPS on a single 4090.Codes are available at (this URL)(https://github.com/HZAI-ZJNU/RemDet).

ASGDiffusion: Parallel High-Resolution Generation with Asynchronous Structure Guidance

Training-free high-resolution (HR) image generation has garnered significant attention due to the high costs of training large diffusion models. Most existing methods begin by reconstructing the overall structure and then proceed to refine the local details. Despite their advancements, they still face issues with repetitive patterns in HR image generation. Besides, HR generation with diffusion models incurs significant computational costs. Thus, parallel generation is essential for interactive applications. To solve the above limitations, we introduce a novel method named ASGDiffusion for parallel HR generation with Asynchronous Structure Guidance (ASG) using pre-trained diffusion models. To solve the pattern repetition problem of HR image generation, ASGDiffusion leverages the low-resolution (LR) noise weighted by the attention mask as the structure guidance for the denoising step to ensure semantic consistency. The proposed structure guidance can significantly alleviate the pattern repetition problem. To enable parallel generation, we further propose a parallelism strategy, which calculates the patch noises and structure guidance asynchronously. By leveraging multi-GPU parallel acceleration, we significantly accelerate generation speed and reduce memory usage per GPU. Extensive experiments demonstrate that our method effectively and efficiently addresses common issues like pattern repetition and achieves state-of-the-art HR generation.

Representation Purification for End-to-End Speech Translation

Speech-to-text translation (ST) is a cross-modal task that involves converting spoken language into text in a different language. Previous research primarily focused on enhancing speech translation by facilitating knowledge transfer from machine translation, exploring various methods to bridge the gap between speech and text modalities. Despite substantial progress made, factors in speech that are not relevant to translation content, such as timbre and rhythm, often limit the efficiency of knowledge transfer. In this paper, we conceptualize speech representation as a combination of content-agnostic and content-relevant factors. We examine the impact of content-agnostic factors on translation performance through preliminary experiments and observe a significant performance deterioration when content-agnostic perturbations are introduced to speech signals. To address this issue, we propose a \textbf{S}peech \textbf{R}epresentation \textbf{P}urification with \textbf{S}upervision \textbf{E}nhancement (SRPSE) framework, which excludes the content-agnostic components within speech representations to mitigate their negative impact on ST. Experiments on MuST-C and CoVoST-2 datasets demonstrate that SRPSE significantly improves translation performance across all translation directions in three settings and achieves preeminent performance under a \textit{transcript-free} setting.

AlignFormer: Modality Matching Can Achieve Better Zero-shot Instruction-Following Speech-LLM

Integrating speech into LLM (speech-LLM) has gaining increased attention recently. The mainstream solution is to connect a well-trained speech encoder and LLM with a neural adapter. However, the length mismatch between the speech and text sequences are not well handled, leading to imperfect modality matching between the speech and text. In this work, we propose a novel neural adapter, AlignFormer, to reduce the length gap between the two modalities. AlignFormer consists of CTC and dynamic-window QFormer layers, where the CTC alignment provides the dynamic window information for qformer layers. The LLM backbone is frozen in training to preserve its text capability, especially the instruction following capability. When training with only the ASR data, the proposed AlignFormer unlocks the instruction following capability for speech-LLM and the model can perform zero-shot speech translation (ST) and speech question answering (SQA) tasks. In fact, speech-LLM with AlignFormer can theoretically perform any tasks that the LLM backbone can deal with in the speech version. To evaluate the effectiveness of the instruction-following speech-LLM, we propose to use instruction following rate (IFR) and offer a systematic perspective for the IFR evaluation. In addition, we find that the audio position in training would affect the instruction following capability of speech-LLM and conduct an in-depth study on it. Our findings show that audio-first training achieves higher IFR than instruction-first training. The AlignFormer can achieve a near 100% IFR with audio-first training and game-changing improvements from zero to non-zero IFR on some evaluation data with instruction-first training. We believe that this study is a big step towards the perfect speech and text modality matching in the LLM embedding space.

Training a Label-Noise-Resistant GNN with Reduced Complexity

Graph Neural Networks (GNNs) have been widely employed for semi-supervised node classification tasks on graphs. However, the performance of GNNs is significantly affected by label noise, that is, a small amount of incorrectly labeled nodes can substantially misguide model training. Mainstream solutions define node classification with label noise (NCLN) as a reliable labeling task, often introducing node similarity with quadratic computational complexity to more accurately assess label reliability. To this end, in this paper, we introduce the Label Ensemble Graph Neural Network (LEGNN), a lower complexity method for robust GNNs training against label noise. LEGNN reframes NCLN as a label ensemble task, gathering informative multiple labels instead of constructing a single reliable label, avoiding high-complexity computations for reliability assessment. Specifically, LEGNN conducts a two-step process: bootstrapping neighboring contexts and robust learning with gathered multiple labels. In the former step, we apply random neighbor masks for each node and gather the predicted labels as a high-probability label set. This mitigates the impact of inaccurately labeled neighbors and diversifies the label set. In the latter step, we utilize a partial label learning based strategy to aggregate the high-probability label information for model training. Additionally, we symmetrically gather a low-probability label set to counteract potential noise from the bootstrapped high-probability label set. Extensive experiments on six datasets demonstrate that LEGNN achieves outstanding performance while ensuring efficiency. Moreover, it exhibits good scalability on dataset with over one hundred thousand nodes and one million edges.

PUMA: Empowering Unified MLLM with Multi-granular Visual Generation

Recent advancements in multimodal foundation models have yielded significant progress in vision-language understanding. Initial attempts have also explored the potential of multimodal large language models (MLLMs) for visual content generation. However, existing works have insufficiently addressed the varying granularity demands of different image generation tasks within a unified MLLM paradigm - from the diversity required in text-to-image generation to the precise controllability needed in image manipulation. In this work, we propose PUMA, emPowering Unified MLLM with Multi-grAnular visual generation. PUMA unifies multi-granular visual features as both inputs and outputs of MLLMs, elegantly addressing the different granularity requirements of various image generation tasks within a unified MLLM framework. Following multimodal pretraining and task-specific instruction tuning, PUMA demonstrates proficiency in a wide range of multimodal tasks. This work represents a significant step towards a truly unified MLLM capable of adapting to the granularity demands of various visual tasks. The code and model will be released in https://github.com/rongyaofang/PUMA.

DreamVideo-2: Zero-Shot Subject-Driven Video Customization with Precise Motion Control

Recent advances in customized video generation have enabled users to create videos tailored to both specific subjects and motion trajectories. However, existing methods often require complicated test-time fine-tuning and struggle with balancing subject learning and motion control, limiting their real-world applications. In this paper, we present DreamVideo-2, a zero-shot video customization framework capable of generating videos with a specific subject and motion trajectory, guided by a single image and a bounding box sequence, respectively, and without the need for test-time fine-tuning. Specifically, we introduce reference attention, which leverages the model's inherent capabilities for subject learning, and devise a mask-guided motion module to achieve precise motion control by fully utilizing the robust motion signal of box masks derived from bounding boxes. While these two components achieve their intended functions, we empirically observe that motion control tends to dominate over subject learning. To address this, we propose two key designs: 1) the masked reference attention, which integrates a blended latent mask modeling scheme into reference attention to enhance subject representations at the desired positions, and 2) a reweighted diffusion loss, which differentiates the contributions of regions inside and outside the bounding boxes to ensure a balance between subject and motion control. Extensive experimental results on a newly curated dataset demonstrate that DreamVideo-2 outperforms state-of-the-art methods in both subject customization and motion control. The dataset, code, and models will be made publicly available.

EvolveDirector: Approaching Advanced Text-to-Image Generation with Large Vision-Language Models

Recent advancements in generation models have showcased remarkable capabilities in generating fantastic content. However, most of them are trained on proprietary high-quality data, and some models withhold their parameters and only provide accessible application programming interfaces (APIs), limiting their benefits for downstream tasks. To explore the feasibility of training a text-to-image generation model comparable to advanced models using publicly available resources, we introduce EvolveDirector. This framework interacts with advanced models through their public APIs to obtain text-image data pairs to train a base model. Our experiments with extensive data indicate that the model trained on generated data of the advanced model can approximate its generation capability. However, it requires large-scale samples of 10 million or more. This incurs significant expenses in time, computational resources, and especially the costs associated with calling fee-based APIs. To address this problem, we leverage pre-trained large vision-language models (VLMs) to guide the evolution of the base model. VLM continuously evaluates the base model during training and dynamically updates and refines the training dataset by the discrimination, expansion, deletion, and mutation operations. Experimental results show that this paradigm significantly reduces the required data volume. Furthermore, when approaching multiple advanced models, EvolveDirector can select the best samples generated by them to learn powerful and balanced abilities. The final trained model Edgen is demonstrated to outperform these advanced models. The code and model weights are available at https://github.com/showlab/EvolveDirector.

CTC-GMM: CTC guided modality matching for fast and accurate streaming speech translation

Models for streaming speech translation (ST) can achieve high accuracy and low latency if they're developed with vast amounts of paired audio in the source language and written text in the target language. Yet, these text labels for the target language are often pseudo labels due to the prohibitive cost of manual ST data labeling. In this paper, we introduce a methodology named Connectionist Temporal Classification guided modality matching (CTC-GMM) that enhances the streaming ST model by leveraging extensive machine translation (MT) text data. This technique employs CTC to compress the speech sequence into a compact embedding sequence that matches the corresponding text sequence, allowing us to utilize matched {source-target} language text pairs from the MT corpora to refine the streaming ST model further. Our evaluations with FLEURS and CoVoST2 show that the CTC-GMM approach can increase translation accuracy relatively by 13.9% and 6.4% respectively, while also boosting decoding speed by 59.7% on GPU.

Hybrid Mamba for Few-Shot Segmentation

Many few-shot segmentation (FSS) methods use cross attention to fuse support foreground (FG) into query features, regardless of the quadratic complexity. A recent advance Mamba can also well capture intra-sequence dependencies, yet the complexity is only linear. Hence, we aim to devise a cross (attention-like) Mamba to capture inter-sequence dependencies for FSS. A simple idea is to scan on support features to selectively compress them into the hidden state, which is then used as the initial hidden state to sequentially scan query features. Nevertheless, it suffers from (1) support forgetting issue: query features will also gradually be compressed when scanning on them, so the support features in hidden state keep reducing, and many query pixels cannot fuse sufficient support features; (2) intra-class gap issue: query FG is essentially more similar to itself rather than to support FG, i.e., query may prefer not to fuse support features but their own ones from the hidden state, yet the success of FSS relies on the effective use of support information. To tackle them, we design a hybrid Mamba network (HMNet), including (1) a support recapped Mamba to periodically recap the support features when scanning query, so the hidden state can always contain rich support information; (2) a query intercepted Mamba to forbid the mutual interactions among query pixels, and encourage them to fuse more support features from the hidden state. Consequently, the support information is better utilized, leading to better performance. Extensive experiments have been conducted on two public benchmarks, showing the superiority of HMNet. The code is available at https://github.com/Sam1224/HMNet.