Duke University
AI Insights
- Marginal revenue: The additional revenue generated by an advertiser for a given slot. [3]
- Virtual value: The derivative of the scaled revenue curve, representing the advertiser's normalized marginal revenue. [3]
- Alaei et al. [3]
- The proposed mechanism uses a new approach called the marginal revenue framework, which focuses on maximizing the sum of marginal revenues across all slots. [3]
- The marginal revenue framework provides a new approach to allocating slots in online advertising auctions, focusing on maximizing the sum of marginal revenues across all slots. [2]
Abstract
Digital advertising platforms and publishers sell ad inventory that conveys targeting information, such as demographic, contextual, or behavioral audience segments, to advertisers. While revealing this information improves ad relevance, it can reduce competition and lower auction revenues. To resolve this trade-off, this paper develops a general auction mechanism -- the Information-Bundling Position Auction (IBPA) mechanism -- that leverages the targeting information to maximize publisher revenue across both search and display advertising environments. The proposed mechanism treats the ad inventory type as the publisher's private information and allocates impressions by comparing advertisers' marginal revenues. We show that IBPA resolves the trade-off between targeting precision and market thickness: publisher revenue is increasing in information granularity and decreasing in disclosure granularity. Moreover, IBPA dominates the generalized second-price (GSP) auction for any distribution of advertiser valuations and under any information or disclosure regime. We also characterize computationally efficient approximations that preserve these guarantees. Using auction-level data from a large retail media platform, we estimate advertiser valuation distributions and simulate counterfactual outcomes. Relative to GSP, IBPA increases publisher revenue by 68%, allocation rate by 19pp, advertiser welfare by 29%, and total welfare by 54%.
Digital advertising platforms and publishers sell ad inventory that conveys targeting information, such as demographic, contextual, or behavioral audience segments, to advertisers. While revealing this information improves ad relevance, it can reduce competition and lower auction revenues. To resolve this trade-off, this paper develops a general auction mechanism -- the Information-Bundling Position Auction (IBPA) mechanism -- that leverages the targeting information to maximize publisher revenue across both search and display advertising environments. The proposed mechanism treats the ad inventory type as the publisher's private information and allocates impressions by comparing advertisers' marginal revenues. We show that IBPA resolves the trade-off between targeting precision and market thickness: publisher revenue is increasing in information granularity and decreasing in disclosure granularity. Moreover, IBPA dominates the generalized second-price (GSP) auction for any distribution of advertiser valuations and under any information or disclosure regime. We also characterize computationally efficient approximations that preserve these guarantees. Using auction-level data from a large retail media platform, we estimate advertiser valuation distributions and simulate counterfactual outcomes. Relative to GSP, IBPA increases publisher revenue by 68%, allocation rate by 19pp, advertiser welfare by 29%, and total welfare by 54%.
Why we are recommending this paper?
Due to your Interest in Bidding
This paper directly addresses the optimization of paid search campaigns by examining targeting information within ad auction mechanisms. Understanding how this information impacts revenue is crucial for maximizing the effectiveness of marketing channels.
Due to your Interest in Bidding
This paper directly addresses the optimization of paid search campaigns by examining targeting information within ad auction mechanisms. Understanding how this information impacts revenue is crucial for maximizing the effectiveness of marketing channels.
NVIDIA
AI Insights
- The paper discusses various methods for attributing model behavior in differentiable games. [3]
- It highlights the importance of understanding how models learn concepts through concept-level attribution. [3]
- They emphasize the need for benchmarking and improving video diffusion transformers for motion transfer. [3]
- Differentiable games: A type of game where the model's behavior can be attributed to specific inputs or actions. [3]
- Concept-level attribution: A method of understanding how models learn concepts through attributing their behavior to specific inputs or actions. [3]
- The paper highlights the importance of understanding model behavior and its applications in differentiable games. [3]
- It emphasizes the need for benchmarking and improving video diffusion transformers for motion transfer. [3]
- The authors also discuss scalable nested optimization as a key technique for efficient training of deep learning models. [3]
- The authors also discuss scalable nested optimization for deep learning and its applications. [2]
Abstract
Despite the rapid progress of video generation models, the role of data in influencing motion is poorly understood. We present Motive (MOTIon attribution for Video gEneration), a motion-centric, gradient-based data attribution framework that scales to modern, large, high-quality video datasets and models. We use this to study which fine-tuning clips improve or degrade temporal dynamics. Motive isolates temporal dynamics from static appearance via motion-weighted loss masks, yielding efficient and scalable motion-specific influence computation. On text-to-video models, Motive identifies clips that strongly affect motion and guides data curation that improves temporal consistency and physical plausibility. With Motive-selected high-influence data, our method improves both motion smoothness and dynamic degree on VBench, achieving a 74.1% human preference win rate compared with the pretrained base model. To our knowledge, this is the first framework to attribute motion rather than visual appearance in video generative models and to use it to curate fine-tuning data.
Despite the rapid progress of video generation models, the role of data in influencing motion is poorly understood. We present Motive (MOTIon attribution for Video gEneration), a motion-centric, gradient-based data attribution framework that scales to modern, large, high-quality video datasets and models. We use this to study which fine-tuning clips improve or degrade temporal dynamics. Motive isolates temporal dynamics from static appearance via motion-weighted loss masks, yielding efficient and scalable motion-specific influence computation. On text-to-video models, Motive identifies clips that strongly affect motion and guides data curation that improves temporal consistency and physical plausibility. With Motive-selected high-influence data, our method improves both motion smoothness and dynamic degree on VBench, achieving a 74.1% human preference win rate compared with the pretrained base model. To our knowledge, this is the first framework to attribute motion rather than visual appearance in video generative models and to use it to curate fine-tuning data.
Why we are recommending this paper?
Due to your Interest in Attribution
Given your interest in personalization and CRM optimization, this NVIDIA research into motion attribution for video generation is highly relevant. Analyzing how data influences video content can inform strategies for creating more engaging and targeted visual experiences.
Due to your Interest in Attribution
Given your interest in personalization and CRM optimization, this NVIDIA research into motion attribution for video generation is highly relevant. Analyzing how data influences video content can inform strategies for creating more engaging and targeted visual experiences.
PingCAP
AI Insights
- HDC generation involves extracting representative entities for each database to facilitate efficient data exploration across multiple databases. [3]
- It also includes a self-refinement chain to correct errors in generated SQL statements. [3]
- The system demonstrates its capabilities through two real-world scenarios: the Financial dataset and the Bird dataset, showcasing its ability to provide insights and facilitate user-system interaction. [3]
- HDC: Hierarchical Data Context - a summary of the data that includes a description, keywords, table information, and more. [3]
- TiChart: Chart Selection - a component that selects the most suitable chart type to present analysis results by visualization. [3]
- Exploration Efficiency: The ability of the system to efficiently explore data across multiple databases. [3]
- TiInsight is a SQL-based automated cross-domain exploratory data analysis system that utilizes large language models to facilitate user-system interaction and provide powerful hierarchical data context (HDC) generation, text-to-SQL (TiSQL), chart selection (TiChart), and exploration efficiency. [2]
- TiSQL is a schema filtering framework based on the map-reduce paradigm that filters tables and columns using clarified questions and cosine similarity. [1]
Abstract
The SQL-based exploratory data analysis has garnered significant attention within the data analysis community. The emergence of large language models (LLMs) has facilitated the paradigm shift from manual to automated data exploration. However, existing methods generally lack the ability for cross-domain analysis, and the exploration of LLMs capabilities remains insufficient. This paper presents TiInsight, an SQL-based automated cross-domain exploratory data analysis system. First, TiInsight offers a user-friendly GUI enabling users to explore data using natural language queries. Second, TiInsight offers a robust cross-domain exploratory data analysis pipeline: hierarchical data context (i.e., HDC) generation, question clarification and decomposition, text-to-SQL (i.e., TiSQL), and data visualization (i.e., TiChart). Third, we have implemented and deployed TiInsight in the production environment of PingCAP and demonstrated its capabilities using representative datasets. The demo video is available at https://youtu.be/JzYFyYd-emI.
The SQL-based exploratory data analysis has garnered significant attention within the data analysis community. The emergence of large language models (LLMs) has facilitated the paradigm shift from manual to automated data exploration. However, existing methods generally lack the ability for cross-domain analysis, and the exploration of LLMs capabilities remains insufficient. This paper presents TiInsight, an SQL-based automated cross-domain exploratory data analysis system. First, TiInsight offers a user-friendly GUI enabling users to explore data using natural language queries. Second, TiInsight offers a robust cross-domain exploratory data analysis pipeline: hierarchical data context (i.e., HDC) generation, question clarification and decomposition, text-to-SQL (i.e., TiSQL), and data visualization (i.e., TiChart). Third, we have implemented and deployed TiInsight in the production environment of PingCAP and demonstrated its capabilities using representative datasets. The demo video is available at https://youtu.be/JzYFyYd-emI.
Why we are recommending this paper?
Due to your Interest in Data Science Management
This work on automated exploratory data analysis using LLMs aligns with your interest in data science management and potentially improving data science organization processes. The SQL-based approach offers a structured way to uncover insights from data.
Due to your Interest in Data Science Management
This work on automated exploratory data analysis using LLMs aligns with your interest in data science management and potentially improving data science organization processes. The SQL-based approach offers a structured way to uncover insights from data.
Fudan University
AI Insights
- On objective tasks, personalized information may not always improve model performance and can even lead to factual errors or logical biases. [3]
- On subjective personalized tasks, personalized information is crucial for improving model performance. [3]
- The type of personalized information used can significantly impact model performance. [3]
- Aligned personas tend to perform better than unaligned personas on both objective and subjective tasks. [3]
- To maximize the benefits of personalized information, it is essential to carefully design and implement the personaDual framework, taking into account the specific requirements and constraints of each application domain. [3]
- Personalized information can have a double-edged effect on the performance of large language models (LLMs). [2]
Abstract
As users increasingly expect LLMs to align with their preferences, personalized information becomes valuable. However, personalized information can be a double-edged sword: it can improve interaction but may compromise objectivity and factual correctness, especially when it is misaligned with the question. To alleviate this problem, we propose PersonaDual, a framework that supports both general-purpose objective reasoning and personalized reasoning in a single model, and adaptively switches modes based on context. PersonaDual is first trained with SFT to learn two reasoning patterns, and then further optimized via reinforcement learning with our proposed DualGRPO to improve mode selection. Experiments on objective and personalized benchmarks show that PersonaDual preserves the benefits of personalization while reducing interference, achieving near interference-free performance and better leveraging helpful personalized signals to improve objective problem-solving.
As users increasingly expect LLMs to align with their preferences, personalized information becomes valuable. However, personalized information can be a double-edged sword: it can improve interaction but may compromise objectivity and factual correctness, especially when it is misaligned with the question. To alleviate this problem, we propose PersonaDual, a framework that supports both general-purpose objective reasoning and personalized reasoning in a single model, and adaptively switches modes based on context. PersonaDual is first trained with SFT to learn two reasoning patterns, and then further optimized via reinforcement learning with our proposed DualGRPO to improve mode selection. Experiments on objective and personalized benchmarks show that PersonaDual preserves the benefits of personalization while reducing interference, achieving near interference-free performance and better leveraging helpful personalized signals to improve objective problem-solving.
Why we are recommending this paper?
Due to your Interest in Personalization
This research directly tackles the challenge of personalization while maintaining objectivity, a key concern for effective CRM and marketing channel strategies. The adaptive reasoning approach is particularly valuable for understanding evolving user preferences.
Due to your Interest in Personalization
This research directly tackles the challenge of personalization while maintaining objectivity, a key concern for effective CRM and marketing channel strategies. The adaptive reasoning approach is particularly valuable for understanding evolving user preferences.
Yonsei University
AI Insights
- Replay buffer: A memory buffer in SPRING that stores a subset of the user's interactions, used to update the parametric adapter. [3]
- Drift score: A measure of how much a user's interaction history has changed over time, used to select samples for the replay buffer. [3]
- ROUGE scores: Metrics used to evaluate the quality of generated text, including ROUGE-1 and ROUGE-L. [3]
- The paper presents SPRING, a framework for adapting large language models (LLMs) to individual users based on their interaction history. [2]
- SPRING: A framework for adapting large language models to individual users based on their interaction history. [1]
Abstract
Personalizing Large Language Models typically relies on static retrieval or one-time adaptation, assuming user preferences remain invariant over time. However, real-world interactions are dynamic, where user interests continuously evolve, posing a challenge for models to adapt to preference drift without catastrophic forgetting. Standard continual learning approaches often struggle in this context, as they indiscriminately update on noisy interaction streams, failing to distinguish genuine preference shifts from transient contexts. To address this, we introduce SPRInG, a novel semi-parametric framework designed for effective continual personalization. During training, SPRInG employs drift-driven selective adaptation, which utilizes a likelihood-based scoring function to identify high-novelty interactions. This allows the model to selectively update the user-specific adapter on drift signals while preserving hard-to-learn residuals in a replay buffer. During inference, we apply strict relevance gating and fuse parametric knowledge with retrieved history via logit interpolation. Experiments on the long-form personalized generation benchmark demonstrate that SPRInG outperforms existing baselines, validating its robustness for real-world continual personalization.
Personalizing Large Language Models typically relies on static retrieval or one-time adaptation, assuming user preferences remain invariant over time. However, real-world interactions are dynamic, where user interests continuously evolve, posing a challenge for models to adapt to preference drift without catastrophic forgetting. Standard continual learning approaches often struggle in this context, as they indiscriminately update on noisy interaction streams, failing to distinguish genuine preference shifts from transient contexts. To address this, we introduce SPRInG, a novel semi-parametric framework designed for effective continual personalization. During training, SPRInG employs drift-driven selective adaptation, which utilizes a likelihood-based scoring function to identify high-novelty interactions. This allows the model to selectively update the user-specific adapter on drift signals while preserving hard-to-learn residuals in a replay buffer. During inference, we apply strict relevance gating and fuse parametric knowledge with retrieved history via logit interpolation. Experiments on the long-form personalized generation benchmark demonstrate that SPRInG outperforms existing baselines, validating its robustness for real-world continual personalization.
Why we are recommending this paper?
Due to your Interest in Personalization
Given your interest in personalization, this paperās focus on continual LLM adaptation to preference drift is highly relevant. The retrieval-interpolated generation approach offers a dynamic solution for optimizing personalization strategies over time.
Due to your Interest in Personalization
Given your interest in personalization, this paperās focus on continual LLM adaptation to preference drift is highly relevant. The retrieval-interpolated generation approach offers a dynamic solution for optimizing personalization strategies over time.
The Hong Kong University of Science and Technology
AI Insights
- The paper discusses the benefits of using mixture-of-experts (MoE) models in natural language processing (NLP). [3]
- MoE models are a type of neural network that consists of multiple experts, each of which is responsible for a specific task or function. [3]
- Mixture-of-experts (MoE) models: A type of neural network that consists of multiple experts, each of which is responsible for a specific task or function. [3]
- Neuron-level knowledge attribution: The process of identifying which neurons in the model are contributing to a particular output or decision. [3]
- Increased complexity and difficulty in interpreting the results Requires careful tuning of hyperparameters to achieve optimal performance Recent studies have shown that MoE models can outperform traditional neural networks on certain tasks, such as language modeling and machine translation. [3]
- However, MoE models also require careful tuning of hyperparameters to achieve optimal performance. [3]
- The authors argue that MoE models can improve performance and efficiency by allowing the model to focus on relevant tasks and ignore irrelevant ones. [2]
Abstract
Mixture-of-Experts (MoE) architectures decouple model capacity from per-token computation, enabling scaling beyond the computational limits imposed by dense scaling laws. Yet how MoE architectures shape knowledge acquisition during pre-training, and how this process differs from dense architectures, remains unknown. To address this issue, we introduce Gated-LPI (Log-Probability Increase), a neuron-level attribution metric that decomposes log-probability increase across neurons. We present a time-resolved comparison of knowledge acquisition dynamics in MoE and dense architectures, tracking checkpoints over 1.2M training steps (~ 5.0T tokens) and 600K training steps (~ 2.5T tokens), respectively. Our experiments uncover three patterns: (1) Low-entropy backbone. The top approximately 1% of MoE neurons capture over 45% of positive updates, forming a high-utility core, which is absent in the dense baseline. (2) Early consolidation. The MoE model locks into a stable importance profile within < 100K steps, whereas the dense model remains volatile throughout training. (3) Functional robustness. Masking the ten most important MoE attention heads reduces relational HIT@10 by < 10%, compared with > 50% for the dense model, showing that sparsity fosters distributed -- rather than brittle -- knowledge storage. These patterns collectively demonstrate that sparsity fosters an intrinsically stable and distributed computational backbone from early in training, helping bridge the gap between sparse architectures and training-time interpretability.
Mixture-of-Experts (MoE) architectures decouple model capacity from per-token computation, enabling scaling beyond the computational limits imposed by dense scaling laws. Yet how MoE architectures shape knowledge acquisition during pre-training, and how this process differs from dense architectures, remains unknown. To address this issue, we introduce Gated-LPI (Log-Probability Increase), a neuron-level attribution metric that decomposes log-probability increase across neurons. We present a time-resolved comparison of knowledge acquisition dynamics in MoE and dense architectures, tracking checkpoints over 1.2M training steps (~ 5.0T tokens) and 600K training steps (~ 2.5T tokens), respectively. Our experiments uncover three patterns: (1) Low-entropy backbone. The top approximately 1% of MoE neurons capture over 45% of positive updates, forming a high-utility core, which is absent in the dense baseline. (2) Early consolidation. The MoE model locks into a stable importance profile within < 100K steps, whereas the dense model remains volatile throughout training. (3) Functional robustness. Masking the ten most important MoE attention heads reduces relational HIT@10 by < 10%, compared with > 50% for the dense model, showing that sparsity fosters distributed -- rather than brittle -- knowledge storage. These patterns collectively demonstrate that sparsity fosters an intrinsically stable and distributed computational backbone from early in training, helping bridge the gap between sparse architectures and training-time interpretability.
Why we are recommending this paper?
Due to your Interest in Attribution
Due to your Interest in Attribution
Indian Institute of Technology Hyderabad
AI Insights
- The paper discusses the properties and decoding methods of a specific type of error-correcting code called BiD codes. [3]
- The authors analyze the structure of these codes, their distance properties, and provide efficient decoding algorithms for certain types of BiD codes. [3]
- BiD(m,r1,r2) - A family of linear block codes with parameters m, r1, and r2, where m is the length of the code, r1 is the dimension of the first component code, and r2 is the dimension of the second component code. [3]
- CA(m,w) - The code C_A(m,w) is defined as the direct sum of two codes: BiD(m,w,w) and the repetition code of length m. [3]
- The paper provides efficient decoding methods for certain types of BiD codes. [3]
- The authors show that the fast implementations of the ML and max-log-MAP decoders from [17] can be applied to BiD(m,1,1) and BiD(m,0,1). [3]
- Additionally, the belief propagation decoding algorithm is described for BiD(m,2,2). [3]
- The paper focuses on a specific type of error-correcting code (BiD codes) and may not be applicable to other types of codes. [3]
- The decoding methods described in the paper are efficient but may not be optimal for all scenarios. [3]
- The paper references a previous work on first-order recursive subproduct codes [17]. [3]
- The paper discusses the properties and decoding methods of BiD codes, a specific type of error-correcting code. [3]
- The authors analyze the structure of these codes, their distance properties, and provide efficient decoding algorithms for certain types of BiD codes. [3]
- BiD(m,0,1) = BiD(m,1,1) ā BiD(m,0,0) - This equation represents the decomposition of BiD(m,0,1) into a direct sum of two codes. [1]
Abstract
BiD codes, which are a new family of algebraic codes of length $3^m$, achieve the erasure channel capacity under bit-MAP decoding and offer asymptotically larger minimum distance than Reed-Muller (RM) codes. In this paper we propose fast maximum-likelihood (ML) and max-log-MAP decoders for first-order BiD codes. For second-order codes, we identify their minimum-weight parity checks and ascertain a code property known as 'projection' in the RM coding literature. We use these results to design a belief propagation decoder that performs within 1 dB of ML decoder for block lengths 81 and 243.
BiD codes, which are a new family of algebraic codes of length $3^m$, achieve the erasure channel capacity under bit-MAP decoding and offer asymptotically larger minimum distance than Reed-Muller (RM) codes. In this paper we propose fast maximum-likelihood (ML) and max-log-MAP decoders for first-order BiD codes. For second-order codes, we identify their minimum-weight parity checks and ascertain a code property known as 'projection' in the RM coding literature. We use these results to design a belief propagation decoder that performs within 1 dB of ML decoder for block lengths 81 and 243.
Why we are recommending this paper?
Due to your Interest in Bidding
Due to your Interest in Bidding
Zhejiang University
AI Insights
- Semantic Communication: A new paradigm that focuses on the meaning and context of the information being transmitted, rather than just its bits. [3]
- The proposed approach has shown promising results in terms of compression ratio and reconstruction quality. [3]
- The proposed approach relies heavily on the quality of the LLM used for compression and decompression. [3]
- There are several existing works that have explored the use of neural networks for lossless data compression, but most of them focus on specific types of data such as images or text. [3]
- The LLM is used as a compressor to compress the input data, and then the compressed data is transmitted over the channel. [3]
- At the receiver end, the LLM is used again to decompress the received data. [3]
- But with this new approach, we use a special kind of computer program called a large language model (LLM) to not only compress the picture but also understand its meaning and context. [3]
- This way, when your friend receives the compressed data, they can easily reconstruct the original picture because the LLM has already understood what it's supposed to look like. [3]
- The paper proposes a novel approach to semantic communication using large language models (LLMs) for source-channel coding. [2]
- Imagine you want to send a picture to your friend over the internet. [1]
Abstract
Separate Source-Channel Coding (SSCC) remains attractive for text transmission due to its modularity and compatibility with mature entropy coders and powerful channel codes. However, SSCC often suffers from a pronounced cliff effect in low Signal-to-Noise Ratio (SNR) regimes, where residual bit errors after channel decoding can catastrophically break lossless source decoding, especially for Arithmetic Coding (AC) driven by Large Language Models (LLMs). This paper proposes a receiver-side In-Context Decoding (ICD) framework that enhances SSCC robustness without modifying the transmitter. ICD leverages an Error Correction Code Transformer (ECCT) to obtain bit-wise reliability for the decoded information bits. Based on the context-consistent bitstream, ICD constructs a confidence-ranked candidate pool via reliability-guided bit flipping, samples a compact yet diverse subset of candidates, and applies an LLM-based arithmetic decoder to obtain both reconstructions and sequence-level log-likelihoods. A reliability-likelihood fusion rule then selects the final output. We further provide theoretical guarantees on the stability and convergence of the proposed sampling procedure. Extensive experiments over Additive White Gaussian Noise (AWGN) and Rayleigh fading channels demonstrate consistent gains compared with conventional SSCC baselines and representative Joint Source-Channel Coding (JSCC) schemes.
Separate Source-Channel Coding (SSCC) remains attractive for text transmission due to its modularity and compatibility with mature entropy coders and powerful channel codes. However, SSCC often suffers from a pronounced cliff effect in low Signal-to-Noise Ratio (SNR) regimes, where residual bit errors after channel decoding can catastrophically break lossless source decoding, especially for Arithmetic Coding (AC) driven by Large Language Models (LLMs). This paper proposes a receiver-side In-Context Decoding (ICD) framework that enhances SSCC robustness without modifying the transmitter. ICD leverages an Error Correction Code Transformer (ECCT) to obtain bit-wise reliability for the decoded information bits. Based on the context-consistent bitstream, ICD constructs a confidence-ranked candidate pool via reliability-guided bit flipping, samples a compact yet diverse subset of candidates, and applies an LLM-based arithmetic decoder to obtain both reconstructions and sequence-level log-likelihoods. A reliability-likelihood fusion rule then selects the final output. We further provide theoretical guarantees on the stability and convergence of the proposed sampling procedure. Extensive experiments over Additive White Gaussian Noise (AWGN) and Rayleigh fading channels demonstrate consistent gains compared with conventional SSCC baselines and representative Joint Source-Channel Coding (JSCC) schemes.
Why we are recommending this paper?
Due to your Interest in Marketing Channels
Due to your Interest in Marketing Channels
Beijing Jiaotong University
AI Insights
- The authors propose a multi-dimensional generalization enhancement strategy that combines pre-training and transfer learning to improve the model's ability to generalize across different environments. [3]
- The study highlights the importance of incorporating propagation knowledge into AI-based channel modeling, which can enhance both prediction accuracy and interpretability. [3]
- RMSE: Root Mean Squared Error Saliency maps: visualizations that highlight the most important regions in an image for a particular task or prediction. [3]
- The proposed multi-dimensional generalization enhancement strategy can significantly improve the model's ability to generalize across different environments, making it more robust and reliable in real-world applications. [3]
- Incorporating propagation knowledge into AI-based channel modeling can enhance both prediction accuracy and interpretability, leading to better decision-making and optimization of wireless communication systems. [3]
- The authors do not provide a detailed comparison with other state-of-the-art methods, making it difficult to evaluate the effectiveness of their approach. [3]
- The authors demonstrate the effectiveness of their approach using quantitative metrics such as faithfulness and comprehensiveness, as well as visualization-based methods like feature-importance visualization techniques. [2]
Abstract
This paper proposes a novel paradigm centered on Artificial Intelligence (AI)-empowered propagation channel prediction to address the limitations of traditional channel modeling. We present a comprehensive framework that deeply integrates heterogeneous environmental data and physical propagation knowledge into AI models for site-specific channel prediction, which referred to as channel inference. By leveraging AI to infer site-specific wireless channel states, the proposed paradigm enables accurate prediction of channel characteristics at both link and area levels, capturing spatio-temporal evolution of radio propagation. Some novel strategies to realize the paradigm are introduced and discussed, including AI-native and AI-hybrid inference approaches. This paper also investigates how to enhance model generalization through transfer learning and improve interpretability via explainable AI techniques. Our approach demonstrates significant practical efficacy, achieving an average path loss prediction root mean square error (RMSE) of $\sim$ 4 dB and reducing training time by 60\%-75\%. This new modeling paradigm provides a foundational pathway toward high-fidelity, generalizable, and physically consistent propagation channel prediction for future communication networks.
This paper proposes a novel paradigm centered on Artificial Intelligence (AI)-empowered propagation channel prediction to address the limitations of traditional channel modeling. We present a comprehensive framework that deeply integrates heterogeneous environmental data and physical propagation knowledge into AI models for site-specific channel prediction, which referred to as channel inference. By leveraging AI to infer site-specific wireless channel states, the proposed paradigm enables accurate prediction of channel characteristics at both link and area levels, capturing spatio-temporal evolution of radio propagation. Some novel strategies to realize the paradigm are introduced and discussed, including AI-native and AI-hybrid inference approaches. This paper also investigates how to enhance model generalization through transfer learning and improve interpretability via explainable AI techniques. Our approach demonstrates significant practical efficacy, achieving an average path loss prediction root mean square error (RMSE) of $\sim$ 4 dB and reducing training time by 60\%-75\%. This new modeling paradigm provides a foundational pathway toward high-fidelity, generalizable, and physically consistent propagation channel prediction for future communication networks.
Why we are recommending this paper?
Due to your Interest in Marketing Channels
Due to your Interest in Marketing Channels
RuhrUniversitt Bochum
AI Insights
- The RDMS is designed to be highly usable, with a focus on reducing the time and effort required for researchers to manage their data. [2]
- Usability testing is conducted regularly through online meetings, with participants completing tasks using the RDMS platform and providing feedback. [1]
Abstract
The goal of the Collaborative Research Center 1625 is the establishment of a scientific basis for the atomic-scale understanding and design of multifunctional compositionally complex solid solution surfaces. Next to materials synthesis in form of thin-film materials libraries, various materials characterization and simulations techniques are used to explore the materials data space of the problem. Machine learning and artificial intelligence techniques guide its exploration and navigation. The effective use of the combined heterogeneous data requires more than just a simple research data management plan. Consequently, our research data management system maps different data modalities in different formats and resolutions from different labs to the correct spatial locations on physical samples. Besides a graphical user interface, the system can also be accessed through an application programming interface for reproducible data-driven workflows. It is implemented by a combination of a custom research data management system designed around a relational database, an ontology which builds upon materials science-specific ontologies, and the construction of a Knowledge Graph. Along with the technical solutions of research data management system and lessons learned, first use cases are shown which were not possible (or at least much harder to achieve) without it.
The goal of the Collaborative Research Center 1625 is the establishment of a scientific basis for the atomic-scale understanding and design of multifunctional compositionally complex solid solution surfaces. Next to materials synthesis in form of thin-film materials libraries, various materials characterization and simulations techniques are used to explore the materials data space of the problem. Machine learning and artificial intelligence techniques guide its exploration and navigation. The effective use of the combined heterogeneous data requires more than just a simple research data management plan. Consequently, our research data management system maps different data modalities in different formats and resolutions from different labs to the correct spatial locations on physical samples. Besides a graphical user interface, the system can also be accessed through an application programming interface for reproducible data-driven workflows. It is implemented by a combination of a custom research data management system designed around a relational database, an ontology which builds upon materials science-specific ontologies, and the construction of a Knowledge Graph. Along with the technical solutions of research data management system and lessons learned, first use cases are shown which were not possible (or at least much harder to achieve) without it.
Why we are recommending this paper?
Due to your Interest in Data Science Management
Due to your Interest in Data Science Management
University of Passau
AI Insights
- The authors propose a novel approach to modeling user behavior using a graph neural network (GNN), which allows for more accurate predictions and better personalization. [3]
- Conversational AI: A type of artificial intelligence that enables computers to understand and respond to human language in a conversational manner. [3]
- Graph Neural Network (GNN): A type of neural network designed to handle graph-structured data, which is particularly useful for modeling complex relationships between entities. [3]
- The development of web-based conversational AI systems has the potential to revolutionize the way we interact with technology and access information. [3]
- The paper cites several studies on conversational AI and user behavior modeling, including the use of GNNs for predicting user interactions. [3]
- The system uses a combination of natural language processing (NLP) and machine learning (ML) techniques to understand user queries and provide relevant responses. [2]
- The paper discusses the development of a web-based conversational AI system that can simulate user interactions and search behaviors. [1]
Abstract
A fundamental tension exists between the demand for sophisticated AI assistance in web search and the need for user data privacy. Current centralized models require users to transmit sensitive browsing data to external services, which limits user control. In this paper, we present a browser extension that provides a viable in-browser alternative. We introduce a hybrid architecture that functions entirely on the client side, combining two components: (1) an adaptive probabilistic model that learns a user's behavioral policy from direct feedback, and (2) a Small Language Model (SLM), running in the browser, which is grounded by the probabilistic model to generate context-aware suggestions. To evaluate this approach, we conducted a three-week longitudinal user study with 18 participants. Our results show that this privacy-preserving approach is highly effective at adapting to individual user behavior, leading to measurably improved search efficiency. This work demonstrates that sophisticated AI assistance is achievable without compromising user privacy or data control.
A fundamental tension exists between the demand for sophisticated AI assistance in web search and the need for user data privacy. Current centralized models require users to transmit sensitive browsing data to external services, which limits user control. In this paper, we present a browser extension that provides a viable in-browser alternative. We introduce a hybrid architecture that functions entirely on the client side, combining two components: (1) an adaptive probabilistic model that learns a user's behavioral policy from direct feedback, and (2) a Small Language Model (SLM), running in the browser, which is grounded by the probabilistic model to generate context-aware suggestions. To evaluate this approach, we conducted a three-week longitudinal user study with 18 participants. Our results show that this privacy-preserving approach is highly effective at adapting to individual user behavior, leading to measurably improved search efficiency. This work demonstrates that sophisticated AI assistance is achievable without compromising user privacy or data control.
Why we are recommending this paper?
Due to your Interest in Paid Search
Due to your Interest in Paid Search
Huawei
AI Insights
- The download cost of the proposed scheme is given by D = 1 + 1/(N-1) * (1 - 1 / ((N-1)e^(-Īµ) + 1)^K-2). [3]
- Download cost D: The total amount of data downloaded by the user to obtain the desired information. [3]
- The proposed scheme satisfies the condition of Īµ-leaky (W,S)-privacy, which means that for any two pairs of demand and side information indices (W,S) and (Wā²,Sā²), the ratio Pr[Q[W,S]n = q|W,S] / Pr[Q[W,S]n = q|Wā²,Sā²] can take only three possible values: either 1, or eāĪµ, or eĪµ. [1]
Abstract
This paper investigates the problem of leaky-private Private Information Retrieval with Side Information (L-PIR-SI), which relaxes the requirement of perfect privacy to achieve improved communication efficiency in the presence of side information. While the capacities of PIR-SI under both $W$-privacy and $(W,S)$-privacy have been partially explored, the impact of controlled information leakage in these settings remains unaddressed. We propose a unified probabilistic framework to construct L-PIR-SI schemes where the privacy leakage is quantified by a parameter $\varepsilon$, consistent with differential privacy standards. We characterize the achievable download costs and show that our results generalize several landmark results in the PIR literature: they recover the capacity of PIR-SI when $\varepsilon \to 0$, and reduce to the known bounds for leaky-PIR when side information is absent. This work provides the first look at the trade-offs between leakage, side information, and retrieval efficiency.
This paper investigates the problem of leaky-private Private Information Retrieval with Side Information (L-PIR-SI), which relaxes the requirement of perfect privacy to achieve improved communication efficiency in the presence of side information. While the capacities of PIR-SI under both $W$-privacy and $(W,S)$-privacy have been partially explored, the impact of controlled information leakage in these settings remains unaddressed. We propose a unified probabilistic framework to construct L-PIR-SI schemes where the privacy leakage is quantified by a parameter $\varepsilon$, consistent with differential privacy standards. We characterize the achievable download costs and show that our results generalize several landmark results in the PIR literature: they recover the capacity of PIR-SI when $\varepsilon \to 0$, and reduce to the known bounds for leaky-PIR when side information is absent. This work provides the first look at the trade-offs between leakage, side information, and retrieval efficiency.
Why we are recommending this paper?
Due to your Interest in Paid Search
Due to your Interest in Paid Search
Interests not found
We did not find any papers that match the below interests.
Try other terms also consider if the content exists in arxiv.org.
- Direction on Data Science Organizations
- customer relationship management (crm) optimization
You can edit or add more interests any time.
š¬ Help Shape Our Pricing
We're exploring pricing options to make this project sustainable. Take 3 minutes to share what you'd be willing to pay (if anything). Your input guides our future investment.
Help us improve your experience!
This project is on its early stages your feedback can be pivotal on the future of the project.
Let us know what you think about this week's papers and suggestions!
Give Feedback