A Learnable Agent Collaboration Network Framework for Personalized Multimodal AI Search Engine

Large language models (LLMs) and retrieval-augmented generation (RAG) techniques have revolutionized traditional information access, enabling AI agent to search and summarize information on behalf of users during dynamic dialogues. Despite their potential, current AI search engines exhibit considerable room for improvement in several critical areas. These areas include the support for multimodal information, the delivery of personalized responses, the capability to logically answer complex questions, and the facilitation of more flexible interactions. This paper proposes a novel AI Search Engine framework called the Agent Collaboration Network (ACN). The ACN framework consists of multiple specialized agents working collaboratively, each with distinct roles such as Account Manager, Solution Strategist, Information Manager, and Content Creator. This framework integrates mechanisms for picture content understanding, user profile tracking, and online evolution, enhancing the AI search engine's response quality, personalization, and interactivity. A highlight of the ACN is the introduction of a Reflective Forward Optimization method (RFO), which supports the online synergistic adjustment among agents. This feature endows the ACN with online learning capabilities, ensuring that the system has strong interactive flexibility and can promptly adapt to user feedback. This learning method may also serve as an optimization approach for agent-based systems, potentially influencing other domains of agent applications.

Beyond KAN: Introducing KarSein for Adaptive High-Order Feature Interaction Modeling in CTR Prediction

Modeling feature interactions is crucial for click-through rate (CTR) prediction, particularly when it comes to high-order explicit interactions. Traditional methods struggle with this task because they often predefine a maximum interaction order, which relies heavily on prior knowledge and can limit the model's effectiveness. Additionally, modeling high-order interactions typically leads to increased computational costs. Therefore, the challenge lies in adaptively modeling high-order feature interactions while maintaining efficiency. To address this issue, we introduce Kolmogorov-Arnold Represented Sparse Efficient Interaction Network (KarSein), designed to optimize both predictive accuracy and computational efficiency. We firstly identify limitations of directly applying Kolmogorov-Arnold Networks (KAN) to CTR and then introduce KarSein to overcome these issues. It features a novel architecture that reduces the computational costs of KAN and supports embedding vectors as feature inputs. Additionally, KarSein employs guided symbolic regression to address the challenge of KAN in spontaneously learning multiplicative relationships. Extensive experiments demonstrate KarSein's superior performance, achieving significant predictive accuracy with minimal computational overhead. Furthermore, KarSein maintains strong global explainability while enabling the removal of redundant features, resulting in a sparse network structure. These advantages also position KarSein as a promising method for efficient inference.

PADRe: A Unifying Polynomial Attention Drop-in Replacement for Efficient Vision Transformer

We present Polynomial Attention Drop-in Replacement (PADRe), a novel and unifying framework designed to replace the conventional self-attention mechanism in transformer models. Notably, several recent alternative attention mechanisms, including Hyena, Mamba, SimA, Conv2Former, and Castling-ViT, can be viewed as specific instances of our PADRe framework. PADRe leverages polynomial functions and draws upon established results from approximation theory, enhancing computational efficiency without compromising accuracy. PADRe's key components include multiplicative nonlinearities, which we implement using straightforward, hardware-friendly operations such as Hadamard products, incurring only linear computational and memory costs. PADRe further avoids the need for using complex functions such as Softmax, yet it maintains comparable or superior accuracy compared to traditional self-attention. We assess the effectiveness of PADRe as a drop-in replacement for self-attention across diverse computer vision tasks. These tasks include image classification, image-based 2D object detection, and 3D point cloud object detection. Empirical results demonstrate that PADRe runs significantly faster than the conventional self-attention (11x ~ 43x faster on server GPU and mobile NPU) while maintaining similar accuracy when substituting self-attention in the transformer models.

Enhancing Retrieval and Managing Retrieval: A Four-Module Synergy for Improved Quality and Efficiency in RAG Systems

Retrieval-augmented generation (RAG) techniques leverage the in-context learning capabilities of large language models (LLMs) to produce more accurate and relevant responses. Originating from the simple 'retrieve-then-read' approach, the RAG framework has evolved into a highly flexible and modular paradigm. A critical component, the Query Rewriter module, enhances knowledge retrieval by generating a search-friendly query. This method aligns input questions more closely with the knowledge base. Our research identifies opportunities to enhance the Query Rewriter module to Query Rewriter+ by generating multiple queries to overcome the Information Plateaus associated with a single query and by rewriting questions to eliminate Ambiguity, thereby clarifying the underlying intent. We also find that current RAG systems exhibit issues with Irrelevant Knowledge; to overcome this, we propose the Knowledge Filter. These two modules are both based on the instruction-tuned Gemma-2B model, which together enhance response quality. The final identified issue is Redundant Retrieval; we introduce the Memory Knowledge Reservoir and the Retriever Trigger to solve this. The former supports the dynamic expansion of the RAG system's knowledge base in a parameter-free manner, while the latter optimizes the cost for accessing external knowledge, thereby improving resource utilization and response efficiency. These four RAG modules synergistically improve the response quality and efficiency of the RAG system. The effectiveness of these modules has been validated through experiments and ablation studies across six common QA datasets. The source code can be accessed at https://github.com/Ancientshi/ERM4.

ERAGent: Enhancing Retrieval-Augmented Language Models with Improved Accuracy, Efficiency, and Personalization

Retrieval-augmented generation (RAG) for language models significantly improves language understanding systems. The basic retrieval-then-read pipeline of response generation has evolved into a more extended process due to the integration of various components, sometimes even forming loop structures. Despite its advancements in improving response accuracy, challenges like poor retrieval quality for complex questions that require the search of multifaceted semantic information, inefficiencies in knowledge re-retrieval during long-term serving, and lack of personalized responses persist. Motivated by transcending these limitations, we introduce ERAGent, a cutting-edge framework that embodies an advancement in the RAG area. Our contribution is the introduction of the synergistically operated module: Enhanced Question Rewriter and Knowledge Filter, for better retrieval quality. Retrieval Trigger is incorporated to curtail extraneous external knowledge retrieval without sacrificing response quality. ERAGent also personalizes responses by incorporating a learned user profile. The efficiency and personalization characteristics of ERAGent are supported by the Experiential Learner module which makes the AI assistant being capable of expanding its knowledge and modeling user profile incrementally. Rigorous evaluations across six datasets and three question-answering tasks prove ERAGent's superior accuracy, efficiency, and personalization, emphasizing its potential to advance the RAG field and its applicability in practical systems.

Parameter Hierarchical Optimization for Visible-Infrared Person Re-Identification

Visible-infrared person re-identification (VI-reID) aims at matching cross-modality pedestrian images captured by disjoint visible or infrared cameras. Existing methods alleviate the cross-modality discrepancies via designing different kinds of network architectures. Different from available methods, in this paper, we propose a novel parameter optimizing paradigm, parameter hierarchical optimization (PHO) method, for the task of VI-ReID. It allows part of parameters to be directly optimized without any training, which narrows the search space of parameters and makes the whole network more easier to be trained. Specifically, we first divide the parameters into different types, and then introduce a self-adaptive alignment strategy (SAS) to automatically align the visible and infrared images through transformation. Considering that features in different dimension have varying importance, we develop an auto-weighted alignment learning (AAL) module that can automatically weight features according to their importance. Importantly, in the alignment process of SAS and AAL, all the parameters are immediately optimized with optimization principles rather than training the whole network, which yields a better parameter training manner. Furthermore, we establish the cross-modality consistent learning (CCL) loss to extract discriminative person representations with translation consistency. We provide both theoretical justification and empirical evidence that our proposed PHO method outperform existing VI-reID approaches.

FutureDepth: Learning to Predict the Future Improves Video Depth Estimation

In this paper, we propose a novel video depth estimation approach, FutureDepth, which enables the model to implicitly leverage multi-frame and motion cues to improve depth estimation by making it learn to predict the future at training. More specifically, we propose a future prediction network, F-Net, which takes the features of multiple consecutive frames and is trained to predict multi-frame features one time step ahead iteratively. In this way, F-Net learns the underlying motion and correspondence information, and we incorporate its features into the depth decoding process. Additionally, to enrich the learning of multiframe correspondence cues, we further leverage a reconstruction network, R-Net, which is trained via adaptively masked auto-encoding of multiframe feature volumes. At inference time, both F-Net and R-Net are used to produce queries to work with the depth decoder, as well as a final refinement network. Through extensive experiments on several benchmarks, i.e., NYUDv2, KITTI, DDAD, and Sintel, which cover indoor, driving, and open-domain scenarios, we show that FutureDepth significantly improves upon baseline models, outperforms existing video depth estimation methods, and sets new state-of-the-art (SOTA) accuracy. Furthermore, FutureDepth is more efficient than existing SOTA video depth estimation models and has similar latencies when comparing to monocular models

DeCoTR: Enhancing Depth Completion with 2D and 3D Attentions

In this paper, we introduce a novel approach that harnesses both 2D and 3D attentions to enable highly accurate depth completion without requiring iterative spatial propagations. Specifically, we first enhance a baseline convolutional depth completion model by applying attention to 2D features in the bottleneck and skip connections. This effectively improves the performance of this simple network and sets it on par with the latest, complex transformer-based models. Leveraging the initial depths and features from this network, we uplift the 2D features to form a 3D point cloud and construct a 3D point transformer to process it, allowing the model to explicitly learn and exploit 3D geometric features. In addition, we propose normalization techniques to process the point cloud, which improves learning and leads to better accuracy than directly using point transformers off the shelf. Furthermore, we incorporate global attention on downsampled point cloud features, which enables long-range context while still being computationally feasible. We evaluate our method, DeCoTR, on established depth completion benchmarks, including NYU Depth V2 and KITTI, showcasing that it sets new state-of-the-art performance. We further conduct zero-shot evaluations on ScanNet and DDAD benchmarks and demonstrate that DeCoTR has superior generalizability compared to existing approaches.

Causal Disentanglement for Regulating Social Influence Bias in Social Recommendation

Social recommendation systems face the problem of social influence bias, which can lead to an overemphasis on recommending items that friends have interacted with. Addressing this problem is crucial, and existing methods often rely on techniques such as weight adjustment or leveraging unbiased data to eliminate this bias. However, we argue that not all biases are detrimental, i.e., some items recommended by friends may align with the user's interests. Blindly eliminating such biases could undermine these positive effects, potentially diminishing recommendation accuracy. In this paper, we propose a Causal Disentanglement-based framework for Regulating Social influence Bias in social recommendation, named CDRSB, to improve recommendation performance. From the perspective of causal inference, we find that the user social network could be regarded as a confounder between the user and item embeddings (treatment) and ratings (outcome). Due to the presence of this social network confounder, two paths exist from user and item embeddings to ratings: a non-causal social influence path and a causal interest path. Building upon this insight, we propose a disentangled encoder that focuses on disentangling user and item embeddings into interest and social influence embeddings. Mutual information-based objectives are designed to enhance the distinctiveness of these disentangled embeddings, eliminating redundant information. Additionally, a regulatory decoder that employs a weight calculation module to dynamically learn the weights of social influence embeddings for effectively regulating social influence bias has been designed. Experimental results on four large-scale real-world datasets Ciao, Epinions, Dianping, and Douban book demonstrate the effectiveness of CDRSB compared to state-of-the-art baselines.

MAMo: Leveraging Memory and Attention for Monocular Video Depth Estimation

We propose MAMo, a novel memory and attention frame-work for monocular video depth estimation. MAMo can augment and improve any single-image depth estimation networks into video depth estimation models, enabling them to take advantage of the temporal information to predict more accurate depth. In MAMo, we augment model with memory which aids the depth prediction as the model streams through the video. Specifically, the memory stores learned visual and displacement tokens of the previous time instances. This allows the depth network to cross-reference relevant features from the past when predicting depth on the current frame. We introduce a novel scheme to continuously update the memory, optimizing it to keep tokens that correspond with both the past and the present visual information. We adopt attention-based approach to process memory features where we first learn the spatio-temporal relation among the resultant visual and displacement memory tokens using self-attention module. Further, the output features of self-attention are aggregated with the current visual features through cross-attention. The cross-attended features are finally given to a decoder to predict depth on the current frame. Through extensive experiments on several benchmarks, including KITTI, NYU-Depth V2, and DDAD, we show that MAMo consistently improves monocular depth estimation networks and sets new state-of-the-art (SOTA) accuracy. Notably, our MAMo video depth estimation provides higher accuracy with lower latency, when omparing to SOTA cost-volume-based video depth models.