Jinjin Xu

Jinjin Xu, Liwu Xu, Yuzhe Yang et al.

Recent advancements in multi-modal large language models (MLLMs) have led to substantial improvements in visual understanding, primarily driven by sophisticated modality alignment strategies. However, predominant approaches prioritize global or regional comprehension, with less focus on fine-grained, pixel-level tasks. To address this gap, we introduce u-LLaVA, an innovative unifying multi-task framework that integrates pixel, regional, and global features to refine the perceptual faculties of MLLMs. We commence by leveraging an efficient modality alignment approach, harnessing both image and video datasets to bolster the model's foundational understanding across diverse visual contexts. Subsequently, a joint instruction tuning method with task-specific projectors and decoders for end-to-end downstream training is presented. Furthermore, this work contributes a novel mask-based multi-task dataset comprising 277K samples, crafted to challenge and assess the fine-grained perception capabilities of MLLMs. The overall framework is simple, effective, and achieves state-of-the-art performance across multiple benchmarks. We also make our model, data, and code publicly accessible at https://github.com/OPPOMKLab/u-LLaVA.

Liwu Xu, Jinjin Xu, Yuzhe Yang et al.

Image Aesthetics Assessment (IAA) is a challenging task due to its subjective nature and expensive manual annotations. Recent large-scale vision-language models, such as Contrastive Language-Image Pre-training (CLIP), have shown their promising representation capability for various downstream tasks. However, the application of CLIP to resource-constrained and low-data IAA tasks remains limited. While few attempts to leverage CLIP in IAA have mainly focused on carefully designed prompts, we extend beyond this by allowing models from different domains and with different model sizes to acquire knowledge from CLIP. To achieve this, we propose a unified and flexible two-phase CLIP-based Semi-supervised Knowledge Distillation (CSKD) paradigm, aiming to learn a lightweight IAA model while leveraging CLIP's strong generalization capability. Specifically, CSKD employs a feature alignment strategy to facilitate the distillation of heterogeneous CLIP teacher and IAA student models, effectively transferring valuable features from pre-trained visual representations to two lightweight IAA models, respectively. To efficiently adapt to downstream IAA tasks in a low-data regime, the two strong visual aesthetics learners then conduct distillation with unlabeled examples for refining and transferring the task-specific knowledge collaboratively. Extensive experiments demonstrate that the proposed CSKD achieves state-of-the-art performance on multiple widely used IAA benchmarks. Furthermore, analysis of attention distance and entropy before and after feature alignment shows the effective transfer of CLIP's feature representation to IAA models, which not only provides valuable guidance for the model initialization of IAA but also enhances the aesthetic feature representation of IAA models. Code will be made publicly available.

Jiaxi Li, Lu Yin, Li Shen et al.

Large Language Models (LLMs) have achieved remarkable capabilities, but their immense computational demands during training remain a critical bottleneck for widespread adoption. Low-rank training has received attention in recent years due to its ability to significantly reduce training memory usage. Meanwhile, applying 2:4 structured sparsity to weights and activations to leverage NVIDIA GPU support for 2:4 structured sparse format has become a promising direction. However, existing low-rank methods often leave activation matrices in full-rank, which dominates memory consumption and limits throughput during large-batch training. Furthermore, directly applying sparsity to weights often leads to non-negligible performance degradation. To achieve efficient pre-training of LLMs, this paper proposes ELAS: Efficient pre-training of Low-rank LLMs via 2:4 Activation Sparsity, a novel framework for low-rank models via 2:4 activation sparsity. ELAS applies squared ReLU activation functions to the feed-forward networks in low-rank models and implements 2:4 structured sparsity on the activations after the squared ReLU operation. We evaluated ELAS through pre-training experiments on LLaMA models ranging from 60M to 1B parameters. The results demonstrate that ELAS maintains performance with minimal degradation after applying 2:4 activation sparsity, while achieving training and inference acceleration. Moreover, ELAS reduces activation memory overhead, particularly with large batch sizes. Code is available at ELAS Repo.

Jiaxi Li, Lu Yin, Li Shen et al.

While large language models (LLMs) have achieved remarkable performance across a wide range of tasks, their massive scale incurs prohibitive computational and memory costs for pre-training from scratch. Recent studies have investigated the use of low-rank parameterization as a means of reducing model size and training cost. In this context, sparsity is often employed as a complementary technique to recover important information lost in low-rank compression by capturing salient features in the residual space. However, existing approaches typically combine low-rank and sparse components in a simplistic or ad hoc manner, often resulting in undesirable performance degradation compared to full-rank training. In this paper, we propose \textbf{LO}w-rank and \textbf{S}parse pre-\textbf{T}raining (\textbf{LOST}) for LLMs, a novel method that ingeniously integrates low-rank and sparse structures to enable effective training of LLMs from scratch under strict efficiency constraints. LOST applies singular value decomposition to weight matrices, preserving the dominant low-rank components, while allocating the remaining singular values to construct channel-wise sparse components to complement the expressiveness of low-rank training. We evaluate LOST on LLM pretraining ranging from 60M to 7B parameters. Our experiments show that LOST achieves competitive or superior performance compared to full-rank models, while significantly reducing both memory and compute overhead. Moreover, Code is available at \href{https://github.com/JiaxiLi1/LOST-Low-rank-and-Sparse-Training-for-Large-Language-Models}{LOST Repo}

Fanyi Wang, Peng Liu, Haotian Hu et al.

Research on diffusion model-based video generation has advanced rapidly. However, limitations in object fidelity and generation length hinder its practical applications. Additionally, specific domains like animated wallpapers require seamless looping, where the first and last frames of the video match seamlessly. To address these challenges, this paper proposes LoopAnimate, a novel method for generating videos with consistent start and end frames. To enhance object fidelity, we introduce a framework that decouples multi-level image appearance and textual semantic information. Building upon an image-to-image diffusion model, our approach incorporates both pixel-level and feature-level information from the input image, injecting image appearance and textual semantic embeddings at different positions of the diffusion model. Existing UNet-based video generation models require to input the entire videos during training to encode temporal and positional information at once. However, due to limitations in GPU memory, the number of frames is typically restricted to 16. To address this, this paper proposes a three-stage training strategy with progressively increasing frame numbers and reducing fine-tuning modules. Additionally, we introduce the Temporal E nhanced Motion Module(TEMM) to extend the capacity for encoding temporal and positional information up to 36 frames. The proposed LoopAnimate, which for the first time extends the single-pass generation length of UNet-based video generation models to 35 frames while maintaining high-quality video generation. Experiments demonstrate that LoopAnimate achieves state-of-the-art performance in both objective metrics, such as fidelity and temporal consistency, and subjective evaluation results.

Fanyi Wang, Binzhi Dong, Haotian Hu et al.

Small Vision Language Models (SVLMs) generally refer to models with parameter sizes less than or equal to 2B. Their low cost and power consumption characteristics confer high commercial value. However, their reasoning abilities are limited by the number of parameters. To address this issue, this paper proposes a post-training optimization paradigm called the Incremental Training Strategy to enhance the reasoning ability of SVLMs. Firstly, we constructed a Self-Supervised Chain-of-Thought (COT) Data Construction System, which leverages multiple LVLMs with 7B parameters or more to transform original data into COT data in a self-supervised manner. Our proposed Incremental Training Strategy consists of four stages. Stage 1 injects domain knowledge by performing Supervised Fine-Tuning (SFT) to the pretrained model on the COT data. Stage 2 aligns the COT data format by conducting a small amount of Group Relative Policy Optimization (GRPO) training constrained only by format rewards on the COT data. Stage 3 enhances reasoning ability by applying GRPO training on the COT data with constraints on both format and accuracy rewards. The resulting model shows significant improvement compared to the baseline. Stage 4 addresses the limited capacity of the SVLMs and the weak ability to capture complex patterns by proposing ClipLow GRPO (CLGRPO) to constrain the capture space of the training process. We conducted extensive comparative and ablation experiments on the abstract semantic recognition dataset EMOSet-118K. Experimental results demonstrate that our method significantly improves the reasoning ability of 1B SVLM. Compared to the baseline model fine-tuned on the original data, accuracy increased by 2.77 and recall by 0.69, achieving performance comparable to that of 8B models.

Jinjin Xu, Yaochu Jin, Wenli Du

Data-driven optimization has found many successful applications in the real world and received increased attention in the field of evolutionary optimization. Most existing algorithms assume that the data used for optimization is always available on a central server for construction of surrogates. This assumption, however, may fail to hold when the data must be collected in a distributed way and is subject to privacy restrictions. This paper aims to propose a federated data-driven evolutionary multi-/many-objective optimization algorithm. To this end, we leverage federated learning for surrogate construction so that multiple clients collaboratively train a radial-basis-function-network as the global surrogate. Then a new federated acquisition function is proposed for the central server to approximate the objective values using the global surrogate and estimate the uncertainty level of the approximated objective values based on the local models. The performance of the proposed algorithm is verified on a series of multi/many-objective benchmark problems by comparing it with two state-of-the-art surrogate-assisted multi-objective evolutionary algorithms.

Hangyu Zhu, Jinjin Xu, Shiqing Liu et al.

Federated learning is an emerging distributed machine learning framework for privacy preservation. However, models trained in federated learning usually have worse performance than those trained in the standard centralized learning mode, especially when the training data are not independent and identically distributed (Non-IID) on the local devices. In this survey, we pro-vide a detailed analysis of the influence of Non-IID data on both parametric and non-parametric machine learning models in both horizontal and vertical federated learning. In addition, cur-rent research work on handling challenges of Non-IID data in federated learning are reviewed, and both advantages and disadvantages of these approaches are discussed. Finally, we suggest several future research directions before concluding the paper.

Jinjin Xu, Yaochu Jin, Wenli Du et al.

Data-driven evolutionary optimization has witnessed great success in solving complex real-world optimization problems. However, existing data-driven optimization algorithms require that all data are centrally stored, which is not always practical and may be vulnerable to privacy leakage and security threats if the data must be collected from different devices. To address the above issue, this paper proposes a federated data-driven evolutionary optimization framework that is able to perform data driven optimization when the data is distributed on multiple devices. On the basis of federated learning, a sorted model aggregation method is developed for aggregating local surrogates based on radial-basis-function networks. In addition, a federated surrogate management strategy is suggested by designing an acquisition function that takes into account the information of both the global and local surrogate models. Empirical studies on a set of widely used benchmark functions in the presence of various data distributions demonstrate the effectiveness of the proposed framework.

Jinjin Xu, Wenli Du, Ran Cheng et al.

Learning over massive data stored in different locations is essential in many real-world applications. However, sharing data is full of challenges due to the increasing demands of privacy and security with the growing use of smart mobile devices and IoT devices. Federated learning provides a potential solution to privacy-preserving and secure machine learning, by means of jointly training a global model without uploading data distributed on multiple devices to a central server. However, most existing work on federated learning adopts machine learning models with full-precision weights, and almost all these models contain a large number of redundant parameters that do not need to be transmitted to the server, consuming an excessive amount of communication costs. To address this issue, we propose a federated trained ternary quantization (FTTQ) algorithm, which optimizes the quantized networks on the clients through a self-learning quantization factor. Theoretical proofs of the convergence of quantization factors, unbiasedness of FTTQ, as well as a reduced weight divergence are given. On the basis of FTTQ, we propose a ternary federated averaging protocol (T-FedAvg) to reduce the upstream and downstream communication of federated learning systems. Empirical experiments are conducted to train widely used deep learning models on publicly available datasets, and our results demonstrate that the proposed T-FedAvg is effective in reducing communication costs and can even achieve slightly better performance on non-IID data in contrast to the canonical federated learning algorithms.