Abstract:Advanced Multimodal Large Language Models (MLLMs) struggle with recent Knowledge-based VQA tasks, such as INFOSEEK and Encyclopedic-VQA, due to their limited and frozen knowledge scope, often leading to ambiguous and inaccurate responses. Thus, multimodal Retrieval-Augmented Generation (mRAG) is naturally introduced to provide MLLMs with comprehensive and up-to-date knowledge, effectively expanding the knowledge scope. However, current mRAG methods have inherent drawbacks, including: 1) Performing retrieval even when external knowledge is not needed. 2) Lacking of identification of evidence that supports the query. 3) Increasing model complexity due to additional information filtering modules or rules. To address these shortcomings, we propose a novel generalized framework called \textbf{m}ultimodal \textbf{R}etrieval-\textbf{R}eflection-\textbf{A}ugmented \textbf{G}eneration (mR$^2$AG), which achieves adaptive retrieval and useful information localization to enable answers through two easy-to-implement reflection operations, preventing high model complexity. In mR$^2$AG, Retrieval-Reflection is designed to distinguish different user queries and avoids redundant retrieval calls, and Relevance-Reflection is introduced to guide the MLLM in locating beneficial evidence of the retrieved content and generating answers accordingly. In addition, mR$^2$AG can be integrated into any well-trained MLLM with efficient fine-tuning on the proposed mR$^2$AG Instruction-Tuning dataset (mR$^2$AG-IT). mR$^2$AG significantly outperforms state-of-the-art MLLMs (e.g., GPT-4v/o) and RAG-based MLLMs on INFOSEEK and Encyclopedic-VQA, while maintaining the exceptional capabilities of base MLLMs across a wide range of Visual-dependent tasks.
Abstract:Existing state estimation algorithms for legged robots that rely on proprioceptive sensors often overlook foot slippage and leg deformation in the physical world, leading to large estimation errors. To address this limitation, we propose a comprehensive measurement model that accounts for both foot slippage and variable leg length by analyzing the relative motion between foot contact points and the robot's body center. We show that leg length is an observable quantity, meaning that its value can be explicitly inferred by designing an auxiliary filter. To this end, we introduce a dual estimation framework that iteratively employs a parameter filter to estimate the leg length parameters and a state filter to estimate the robot's state. To prevent error accumulation in this iterative framework, we construct a partial measurement model for the parameter filter using the leg static equation. This approach ensures that leg length estimation relies solely on joint torques and foot contact forces, avoiding the influence of state estimation errors on the parameter estimation. Unlike leg length which can be directly estimated, foot slippage cannot be measured directly with the current sensor configuration. However, since foot slippage occurs at a low frequency, it can be treated as outliers in the measurement data. To mitigate the impact of these outliers, we propose the beta Kalman filter (beta KF), which redefines the estimation loss in canonical Kalman filtering using beta divergence. This divergence can assign low weights to outliers in an adaptive manner, thereby enhancing the robustness of the estimation algorithm. These techniques together form the dual beta-Kalman filter (Dual beta KF), a novel algorithm for robust state estimation in legged robots. Experimental results on the Unitree GO2 robot demonstrate that the Dual beta KF significantly outperforms state-of-the-art methods.
Abstract:Solving constrained multi-objective optimization problems (CMOPs) is a challenging task. While many practical algorithms have been developed to tackle CMOPs, real-world scenarios often present cases where the constraint functions are unknown or unquantifiable, resulting in only binary outcomes (feasible or infeasible). This limitation reduces the effectiveness of constraint violation guidance, which can negatively impact the performance of existing algorithms that rely on this approach. Such challenges are particularly detrimental for algorithms employing the epsilon-based method, as they hinder effective relaxation of the feasible region. To address these challenges, this paper proposes a novel algorithm called DRMCMO based on the detection region method. In DRMCMO, detection regions dynamic monitor feasible solutions to enhance convergence, helping the population escape local optima. Additionally, these regions collaborate with the neighbor pairing strategy to improve population diversity within narrow feasible areas. We have modified three existing test suites to serve as benchmark test problems for CMOPs with binary constraints(CMOP/BC) and conducted comprehensive comparative experiments with state-of-the-art algorithms on these test suites and real-world problems. The results demonstrate the strong competitiveness of DRMCMO against state-of-the-art algorithms. Given the limited research on CMOP/BC, our study offers a new perspective for advancing this field.
Abstract:Artificial Intelligence (AI) systems for online recruitment markets have the potential to significantly enhance the efficiency and effectiveness of job placements and even promote fairness or inclusive hiring practices. Neglecting Diversity and Inclusion (D&I) in these systems, however, can perpetuate biases, leading to unfair hiring practices and decreased workplace diversity, while exposing organisations to legal and reputational risks. Despite the acknowledged importance of D&I in AI, there is a gap in research on effectively implementing D&I guidelines in real-world recruitment systems. Challenges include a lack of awareness and framework for operationalising D&I in a cost-effective, context-sensitive manner. This study aims to investigate the practical application of D&I guidelines in AI-driven online job-seeking systems, specifically exploring how these principles can be operationalised to create more inclusive recruitment processes. We conducted a co-design workshop with a large multinational recruitment company focusing on two AI-driven recruitment use cases. User stories and personas were applied to evaluate the impacts of AI on diverse stakeholders. Follow-up interviews were conducted to assess the workshop's long-term effects on participants' awareness and application of D&I principles. The co-design workshop successfully increased participants' understanding of D&I in AI. However, translating awareness into operational practice posed challenges, particularly in balancing D&I with business goals. The results suggest developing tailored D&I guidelines and ongoing support to ensure the effective adoption of inclusive AI practices.
Abstract:Large language models (LLMs) pre-trained on massive corpora have demonstrated impressive few-shot learning capability on many NLP tasks. Recasting an NLP task into a text-to-text generation task is a common practice so that generative LLMs can be prompted to resolve it. However, performing document-level relation extraction (DocRE) tasks with generative LLM models is still challenging due to the structured output format of DocRE, which complicates the conversion to plain text. Limited information available in few-shot samples and prompt instructions induce further difficulties and challenges in relation extraction for mentioned entities in a document. In this paper, we represent the structured output as a graph-style triplet rather than natural language expressions and leverage generative LLMs for the DocRE task. Our approach, the Graph-DPEP framework is grounded in the reasoning behind triplet explanation thoughts presented in natural language. In this framework, we first introduce a ``decomposed-plug" method for performing the generation from LLMs over prompts with type-space decomposition to alleviate the burden of distinguishing all relation types. Second, we employ a verifier for calibrating the generation and identifying overlooked query entity pairs. Third, we develop "ensemble-play", reapplying generation on the entire type list by leveraging the reasoning thoughts embedded in a sub-graph associated with the missing query pair to address the missingness issue. Through extensive comparisons with existing prompt techniques and alternative Language Models (LLMs), our framework demonstrates superior performance on publicly available benchmarks in experiments.
Abstract:Point cloud is often regarded as a discrete sampling of Riemannian manifold and plays a pivotal role in the 3D image interpretation. Particularly, rotation perturbation, an unexpected small change in rotation caused by various factors (like equipment offset, system instability, measurement errors and so on), can easily lead to the inferior results in point cloud learning tasks. However, classical point cloud learning methods are sensitive to rotation perturbation, and the existing networks with rotation robustness also have much room for improvements in terms of performance and noise tolerance. Given these, this paper remodels the point cloud from the perspective of manifold as well as designs a manifold distillation method to achieve the robustness of rotation perturbation without any coordinate transformation. In brief, during the training phase, we introduce a teacher network to learn the rotation robustness information and transfer this information to the student network through online distillation. In the inference phase, the student network directly utilizes the original 3D coordinate information to achieve the robustness of rotation perturbation. Experiments carried out on four different datasets verify the effectiveness of our method. Averagely, on the Modelnet40 and ScanobjectNN classification datasets with random rotation perturbations, our classification accuracy has respectively improved by 4.92% and 4.41%, compared to popular rotation-robust networks; on the ShapeNet and S3DIS segmentation datasets, compared to the rotation-robust networks, the improvements of mIoU are 7.36% and 4.82%, respectively. Besides, from the experimental results, the proposed algorithm also shows excellent performance in resisting noise and outliers.
Abstract:The emergence of models like GPTs, Claude, LLaMA, and Qwen has reshaped AI applications, presenting vast new opportunities across industries. Yet, the integration of tabular data remains notably underdeveloped, despite its foundational role in numerous real-world domains. This gap is critical for three main reasons. First, database or data warehouse data integration is essential for advanced applications; second, the vast and largely untapped resource of tabular data offers immense potential for analysis; and third, the business intelligence domain specifically demands adaptable, precise solutions that many current LLMs may struggle to provide. In response, we introduce TableGPT2, a model rigorously pre-trained and fine-tuned with over 593.8K tables and 2.36M high-quality query-table-output tuples, a scale of table-related data unprecedented in prior research. This extensive training enables TableGPT2 to excel in table-centric tasks while maintaining strong general language and coding abilities. One of TableGPT2's key innovations is its novel table encoder, specifically designed to capture schema-level and cell-level information. This encoder strengthens the model's ability to handle ambiguous queries, missing column names, and irregular tables commonly encountered in real-world applications. Similar to visual language models, this pioneering approach integrates with the decoder to form a robust large multimodal model. We believe the results are compelling: over 23 benchmarking metrics, TableGPT2 achieves an average performance improvement of 35.20% in the 7B model and 49.32% in the 72B model over prior benchmark-neutral LLMs, with robust general-purpose capabilities intact.
Abstract:Semantic segmentation is an important branch of image processing and computer vision. With the popularity of deep learning, various deep semantic segmentation networks have been proposed for pixel-level classification and segmentation tasks. However, the imaging angles are often arbitrary in real world, such as water body images in remote sensing, and capillary and polyp images in medical field, and we usually cannot obtain prior orientation information to guide these networks to extract more effective features. Additionally, learning the features of objects with multiple orientation information is also challenging, as most CNN-based semantic segmentation networks do not have rotation equivariance to resist the disturbance from orientation information. To address the same, in this paper, we first establish a universal convolution-group framework to more fully utilize the orientation information and make the networks rotation equivariant. Then, we mathematically construct the padding-based rotation equivariant convolution mode (PreCM), which can be used not only for multi-scale images and convolution kernels, but also as a replacement component to replace multiple convolutions, like dilated convolution, transposed convolution, variable stride convolution, etc. In order to verify the realization of rotation equivariance, a new evaluation metric named rotation difference (RD) is finally proposed. The experiments carried out on the datesets Satellite Images of Water Bodies, DRIVE and Floodnet show that the PreCM-based networks can achieve better segmentation performance than the original and data augmentation-based networks. In terms of the average RD value, the former is 0% and the latter two are respectively 7.0503% and 3.2606%. Last but not least, PreCM also effectively enhances the robustness of networks to rotation perturbations.
Abstract:We introduce Baichuan Alignment, a detailed analysis of the alignment techniques employed in the Baichuan series of models. This represents the industry's first comprehensive account of alignment methodologies, offering valuable insights for advancing AI research. We investigate the critical components that enhance model performance during the alignment process, including optimization methods, data strategies, capability enhancements, and evaluation processes. The process spans three key stages: Prompt Augmentation System (PAS), Supervised Fine-Tuning (SFT), and Preference Alignment. The problems encountered, the solutions applied, and the improvements made are thoroughly recorded. Through comparisons across well-established benchmarks, we highlight the technological advancements enabled by Baichuan Alignment. Baichuan-Instruct is an internal model, while Qwen2-Nova-72B and Llama3-PBM-Nova-70B are instruct versions of the Qwen2-72B and Llama-3-70B base models, optimized through Baichuan Alignment. Baichuan-Instruct demonstrates significant improvements in core capabilities, with user experience gains ranging from 17% to 28%, and performs exceptionally well on specialized benchmarks. In open-source benchmark evaluations, both Qwen2-Nova-72B and Llama3-PBM-Nova-70B consistently outperform their respective official instruct versions across nearly all datasets. This report aims to clarify the key technologies behind the alignment process, fostering a deeper understanding within the community. Llama3-PBM-Nova-70B model is available at https://huggingface.co/PKU-Baichuan-MLSystemLab/Llama3-PBM-Nova-70B.
Abstract:The salient multimodal capabilities and interactive experience of GPT-4o highlight its critical role in practical applications, yet it lacks a high-performing open-source counterpart. In this paper, we introduce Baichuan-Omni, the first open-source 7B Multimodal Large Language Model (MLLM) adept at concurrently processing and analyzing modalities of image, video, audio, and text, while delivering an advanced multimodal interactive experience and strong performance. We propose an effective multimodal training schema starting with 7B model and proceeding through two stages of multimodal alignment and multitask fine-tuning across audio, image, video, and text modal. This approach equips the language model with the ability to handle visual and audio data effectively. Demonstrating strong performance across various omni-modal and multimodal benchmarks, we aim for this contribution to serve as a competitive baseline for the open-source community in advancing multimodal understanding and real-time interaction.