Northeastern University
AI Insights
- The paper highlights the importance of understanding and improving information transfer in multi-task learning, which can lead to better performance on downstream tasks. (ML: 0.97)
- Understanding and Improving Information Transfer in Multi-Task Learning Imagine you're trying to learn multiple skills at the same time. (ML: 0.97)
- The paper provides a comprehensive overview of the state-of-the-art in understanding and improving information transfer in multi-task learning. (ML: 0.96)
- Multi-task learning: A machine learning approach that involves training a model on multiple tasks simultaneously to improve its performance on each task. (ML: 0.96)
- The paper is about how to make this process more efficient by understanding how information is transferred between these different tasks. (ML: 0.96)
- Information transfer: The process by which knowledge or information is transferred from one task to another during multi-task learning. (ML: 0.96)
- The paper discusses various approaches to understanding and improving information transfer, including influence functions, gradient-based estimation, and kernel-based methods. (ML: 0.96)
- For example, learning a new language and playing a musical instrument simultaneously. (ML: 0.96)
- Influence function: A mathematical concept used to measure the change in a model's parameters when a new data point is added or removed. (ML: 0.88)
- The paper focuses primarily on theoretical aspects and may not provide practical solutions for real-world applications. (ML: 0.63)
Abstract
Modern AI agents such as large language models are trained on diverse tasks -- translation, code generation, mathematical reasoning, and text prediction -- simultaneously. A key question is to quantify how each individual training task influences performance on a target task, a problem we refer to as task attribution. The direct approach, leave-one-out retraining, measures the effect of removing each task, but is computationally infeasible at scale. An alternative approach that builds surrogate models to predict a target task's performance for any subset of training tasks has emerged in recent literature. Prior work focuses on linear surrogate models, which capture first-order relationships, but miss nonlinear interactions such as synergy, antagonism, or XOR-type effects. In this paper, we first consider a unified task weighting framework for analyzing task attribution methods, and show a new connection between linear surrogate models and influence functions through a second-order analysis. Then, we introduce kernel surrogate models, which more effectively represent second-order task interactions. To efficiently learn the kernel surrogate, we develop a gradient-based estimation procedure that leverages a first-order approximation of pretrained models; empirically, this yields accurate estimates with less than $2\%$ relative error without repeated retraining. Experiments across multiple domains -- including math reasoning in transformers, in-context learning, and multi-objective reinforcement learning -- demonstrate the effectiveness of kernel surrogate models. They achieve a $25\%$ higher correlation with the leave-one-out ground truth than linear surrogates and influence-function baselines. When used for downstream task selection, kernel surrogate models yield a $40\%$ improvement in demonstration selection for in-context learning and multi-objective reinforcement learning benchmarks.
Modern AI agents such as large language models are trained on diverse tasks -- translation, code generation, mathematical reasoning, and text prediction -- simultaneously. A key question is to quantify how each individual training task influences performance on a target task, a problem we refer to as task attribution. The direct approach, leave-one-out retraining, measures the effect of removing each task, but is computationally infeasible at scale. An alternative approach that builds surrogate models to predict a target task's performance for any subset of training tasks has emerged in recent literature. Prior work focuses on linear surrogate models, which capture first-order relationships, but miss nonlinear interactions such as synergy, antagonism, or XOR-type effects. In this paper, we first consider a unified task weighting framework for analyzing task attribution methods, and show a new connection between linear surrogate models and influence functions through a second-order analysis. Then, we introduce kernel surrogate models, which more effectively represent second-order task interactions. To efficiently learn the kernel surrogate, we develop a gradient-based estimation procedure that leverages a first-order approximation of pretrained models; empirically, this yields accurate estimates with less than $2\%$ relative error without repeated retraining. Experiments across multiple domains -- including math reasoning in transformers, in-context learning, and multi-objective reinforcement learning -- demonstrate the effectiveness of kernel surrogate models. They achieve a $25\%$ higher correlation with the leave-one-out ground truth than linear surrogates and influence-function baselines. When used for downstream task selection, kernel surrogate models yield a $40\%$ improvement in demonstration selection for in-context learning and multi-objective reinforcement learning benchmarks.
Why we are recommending this paper?
Due to your Interest in Attribution
This paper directly addresses attribution, a core interest, by exploring how different training tasks influence performance. The focus on AI agents and task quantification aligns strongly with data science management and personalization strategies.
Due to your Interest in Attribution
This paper directly addresses attribution, a core interest, by exploring how different training tasks influence performance. The focus on AI agents and task quantification aligns strongly with data science management and personalization strategies.
Xiamen University
AI Insights
- Fairness metrics: Quantitative measures used to evaluate the fairness of a model's predictions. (ML: 0.99)
- The authors use a combination of fairness metrics, including demographic parity and equalized odds, to evaluate the effectiveness of their approach. (ML: 0.99)
- Debiasing: The process of reducing or eliminating biases in machine learning models, particularly those that affect marginalized groups. (ML: 0.99)
- The use of multi-objective optimization and fairness metrics provides a more comprehensive evaluation of the model's performance. (ML: 0.99)
- The method aims to minimize the bias in the model's predictions while maintaining its overall performance. (ML: 0.97)
- The proposed approach demonstrates promising results in reducing bias in large language models. (ML: 0.96)
- The paper proposes a novel approach to debiasing large language models using a multi-objective optimization framework. (ML: 0.94)
- However, further research is needed to fully understand its limitations and potential applications. (ML: 0.93)
- Multi-objective optimization: A technique used to optimize multiple conflicting objectives simultaneously. (ML: 0.89)
- Limited experimental results (ML: 0.76)
Abstract
Large language models (LLMs) have demonstrated impressive capabilities across a wide range of natural language processing tasks. However, their outputs often exhibit social biases, raising fairness concerns. Existing debiasing methods, such as fine-tuning on additional datasets or prompt engineering, face scalability issues or compromise user experience in multi-turn interactions. To address these challenges, we propose a framework for detecting stereotype-inducing words and attributing neuron-level bias in LLMs, without the need for fine-tuning or prompt modification. Our framework first identifies stereotype-inducing adjectives and nouns via comparative analysis across demographic groups. We then attribute biased behavior to specific neurons using two attribution strategies based on integrated gradients. Finally, we mitigate bias by directly intervening on their activations at the projection layer. Experiments on three widely used LLMs demonstrate that our method effectively reduces bias while preserving overall model performance. Code is available at the github link: https://github.com/XMUDeepLIT/Bi-directional-Bias-Attribution.
Large language models (LLMs) have demonstrated impressive capabilities across a wide range of natural language processing tasks. However, their outputs often exhibit social biases, raising fairness concerns. Existing debiasing methods, such as fine-tuning on additional datasets or prompt engineering, face scalability issues or compromise user experience in multi-turn interactions. To address these challenges, we propose a framework for detecting stereotype-inducing words and attributing neuron-level bias in LLMs, without the need for fine-tuning or prompt modification. Our framework first identifies stereotype-inducing adjectives and nouns via comparative analysis across demographic groups. We then attribute biased behavior to specific neurons using two attribution strategies based on integrated gradients. Finally, we mitigate bias by directly intervening on their activations at the projection layer. Experiments on three widely used LLMs demonstrate that our method effectively reduces bias while preserving overall model performance. Code is available at the github link: https://github.com/XMUDeepLIT/Bi-directional-Bias-Attribution.
Why we are recommending this paper?
Due to your Interest in Attribution
Given the interest in CRM optimization and personalization, this work offers a method for addressing bias in LLMs, a critical concern for responsible AI development. Understanding and mitigating bias is essential for effective personalization strategies.
Due to your Interest in Attribution
Given the interest in CRM optimization and personalization, this work offers a method for addressing bias in LLMs, a critical concern for responsible AI development. Understanding and mitigating bias is essential for effective personalization strategies.
National Taiwan University
AI Insights
- The dataset is used to train a Temporal Graph Neural Network (TGN) model, which generates probability scores for each congressional transaction. (ML: 0.94)
- The existing literature on congressional trading is classified into three categories: historical performance relative to regulation, market-based predictors of outcomes, and performance attributed to specific leadership roles or regime trends. (ML: 0.94)
- Temporal Graph Neural Network (TGN): A type of neural network designed for modeling temporal relationships between entities in a graph. (ML: 0.89)
- The TGN model is then combined with a portfolio allocation model using a Time-Series Stacking Architecture. (ML: 0.88)
- The proposed portfolio allocation model redetermines the portfolio configuration daily by processing prediction labels associated with each trade beginning on the day of the filing. (ML: 0.85)
- 64 macroeconomic indicators are collected to serve as global context features, updated daily based on the most recent value reported by the Federal Reserve Bank of St. (ML: 0.85)
- The performance of five portfolios will be compared using metrics such as Annualized Return and Volatility, Sharpe Ratio, Maximum Drawdown, Sortino and Calmar Ratios. (ML: 0.83)
- The Capitol Gains Dataset is a comprehensive collection of data related to congressional trading, including transactions, lobbying reports, and economic indicators. (ML: 0.82)
- The model incorporates a signal decay mechanism to ensure that older signals exercise decreasing influence on the current allocation. (ML: 0.80)
- Corporate financial facts are extracted from quarterly 10-Q filings via the SEC EDGAR API, indexed by their filing date to prevent lookahead bias. (ML: 0.73)
- Louis (FRED/ALFRED). (ML: 0.64)
Abstract
Congressional stock trading has raised concerns about potential information asymmetries and conflicts of interest in financial markets. We introduce a temporal graph network (TGN) framework to identify information channels through which members of Congress may possess advantageous knowledge when trading company stocks. We construct a multimodal dynamic graph integrating diverse publicly available datasets, including congressional stock transactions, lobbying relationships, campaign finance contributions, and geographical connections between legislators and corporations. Our approach formulates the detection problem as a dynamic edge classification task, where we identify trades that exhibit statistically significant outperformance relative to the S&P 500 across long time horizons. To handle the temporal nature of these relationships, we develop a two-step walk-forward validation architecture that respects information availability constraints and prevents look-ahead bias. We evaluate several labeling strategies based on risk-adjusted returns and demonstrate that the TGN successfully captures complex temporal dependencies between congressional-corporate interactions and subsequent trading performance.
Congressional stock trading has raised concerns about potential information asymmetries and conflicts of interest in financial markets. We introduce a temporal graph network (TGN) framework to identify information channels through which members of Congress may possess advantageous knowledge when trading company stocks. We construct a multimodal dynamic graph integrating diverse publicly available datasets, including congressional stock transactions, lobbying relationships, campaign finance contributions, and geographical connections between legislators and corporations. Our approach formulates the detection problem as a dynamic edge classification task, where we identify trades that exhibit statistically significant outperformance relative to the S&P 500 across long time horizons. To handle the temporal nature of these relationships, we develop a two-step walk-forward validation architecture that respects information availability constraints and prevents look-ahead bias. We evaluate several labeling strategies based on risk-adjusted returns and demonstrate that the TGN successfully captures complex temporal dependencies between congressional-corporate interactions and subsequent trading performance.
Why we are recommending this paper?
Due to your Interest in Marketing Channels
This paper’s investigation into information channels within financial markets is highly relevant to understanding data science management and potentially identifying patterns in paid search campaigns. The use of graph learning is also a valuable data science technique.
Due to your Interest in Marketing Channels
This paper’s investigation into information channels within financial markets is highly relevant to understanding data science management and potentially identifying patterns in paid search campaigns. The use of graph learning is also a valuable data science technique.
Stanford University
AI Insights
- σp,vk and σ′p,wk: population standard deviations of publisher revenue for treatments v and w, respectively. (ML: 0.97)
- q: a parameter representing the proportion of users with treatment v. (ML: 0.97)
- ηv: a parameter representing the expected value of publisher revenue for treatment v. (ML: 0.96)
- Eπr and Eπ′r: population means of publisher revenue for treatments v and w, respectively. (ML: 0.95)
- The problem statement is a mathematical proof of a confidence interval for estimating publisher revenue in an online advertising system. (ML: 0.93)
- Φ(v) and Φ(w): Markov chains generated by large numbers d and d/2 of users. (ML: 0.92)
- The main result is a confidence interval for the estimate of publisher revenue with treatment v in Markov chain Φ(v). (ML: 0.92)
- SKv,k(v,˜Φ(v)) and SWk,w(˜Φ(w)): sample means of publisher revenue for treatment v in Markov chain Φ(v) and treatment w in Markov chain Φ(w). (ML: 0.91)
- The proof involves several lemmas and theorems, including Anscombe's Theorem, which provides a bound on the probability that the sample mean deviates from the population mean by more than a certain amount. (ML: 0.88)
- The proof uses several mathematical techniques, including the Triangle Inequality and the Kolmogorov Inequality. (ML: 0.82)
Abstract
We consider a causal inference problem frequently encountered in online advertising systems, where a publisher (e.g., Instagram, TikTok) interacts repeatedly with human users and advertisers by sporadically displaying to each user an advertisement selected through an auction. Each treatment corresponds to a parameter value of the advertising mechanism (e.g., auction reserve-price), and we want to estimate through experiments the corresponding long-term treatment effect (e.g., annual advertising revenue). In our setting, the treatment affects not only the instantaneous revenue from showing an ad, but also changes each user's interaction-trajectory, and each advertiser's bidding policy -- as the latter is constrained by a finite budget. In particular, each a treatment may even affect the size of the population, since users interact longer with a tolerable advertising mechanism. We drop the classical i.i.d. assumption and model the experiment measurements (e.g., advertising revenue) as a stopped random walk, and use a budget-splitting experimental design, the Anscombe Theorem, a Wald-like equation, and a Central Limit Theorem to construct confidence intervals for the long-term treatment effect.
We consider a causal inference problem frequently encountered in online advertising systems, where a publisher (e.g., Instagram, TikTok) interacts repeatedly with human users and advertisers by sporadically displaying to each user an advertisement selected through an auction. Each treatment corresponds to a parameter value of the advertising mechanism (e.g., auction reserve-price), and we want to estimate through experiments the corresponding long-term treatment effect (e.g., annual advertising revenue). In our setting, the treatment affects not only the instantaneous revenue from showing an ad, but also changes each user's interaction-trajectory, and each advertiser's bidding policy -- as the latter is constrained by a finite budget. In particular, each a treatment may even affect the size of the population, since users interact longer with a tolerable advertising mechanism. We drop the classical i.i.d. assumption and model the experiment measurements (e.g., advertising revenue) as a stopped random walk, and use a budget-splitting experimental design, the Anscombe Theorem, a Wald-like equation, and a Central Limit Theorem to construct confidence intervals for the long-term treatment effect.
Why we are recommending this paper?
Due to your Interest in Marketing Channels
The focus on causal inference within online advertising directly relates to bidding strategies and understanding user behavior. This research offers a framework for optimizing paid search campaigns based on causal relationships.
Due to your Interest in Marketing Channels
The focus on causal inference within online advertising directly relates to bidding strategies and understanding user behavior. This research offers a framework for optimizing paid search campaigns based on causal relationships.
Amazon
AI Insights
- One potential weakness of this approach is that it may be limited by the availability of high-quality training data and the complexity of the problems being addressed. (ML: 0.98)
- With AI, this process becomes even more efficient and accurate. (ML: 0.98)
- Imagine you're trying to solve a complex math problem, but you're not sure where to start. (ML: 0.95)
- The use of AI in formal mathematical reasoning has opened up new possibilities for solving complex problems and making accurate predictions. (ML: 0.93)
- Formal mathematical reasoning has been increasingly used in various fields, including computer science, mathematics, and artificial intelligence. (ML: 0.93)
- Formal mathematical reasoning is like having a super-smart assistant that can help you break down the problem into smaller parts and find the solution. (ML: 0.92)
- The main idea is to explore the intersection of formal mathematical reasoning and AI, highlighting its potential applications and benefits. (ML: 0.88)
- The paper discusses the development and application of formal mathematical reasoning using AI. (ML: 0.88)
Abstract
We introduce CSLib, an open-source framework for proving computer-science-related theorems and writing formally verified code in the Lean proof assistant. CSLib aims to be for computer science what Lean's Mathlib is for mathematics. Mathlib has been tremendously impactful: it is a key reason for Lean's popularity within the mathematics research community, and it has also played a critical role in the training of AI systems for mathematical reasoning. However, the base of computer science knowledge in Lean is currently quite limited. CSLib will vastly enhance this knowledge base and provide infrastructure for using this knowledge in real-world verification projects. By doing so, CSLib will (1) enable the broad use of Lean in computer science education and research, and (2) facilitate the manual and AI-aided engineering of large-scale formally verified systems.
We introduce CSLib, an open-source framework for proving computer-science-related theorems and writing formally verified code in the Lean proof assistant. CSLib aims to be for computer science what Lean's Mathlib is for mathematics. Mathlib has been tremendously impactful: it is a key reason for Lean's popularity within the mathematics research community, and it has also played a critical role in the training of AI systems for mathematical reasoning. However, the base of computer science knowledge in Lean is currently quite limited. CSLib will vastly enhance this knowledge base and provide infrastructure for using this knowledge in real-world verification projects. By doing so, CSLib will (1) enable the broad use of Lean in computer science education and research, and (2) facilitate the manual and AI-aided engineering of large-scale formally verified systems.
Why we are recommending this paper?
Due to your Interest in Data Science Management
Coming from Amazon, this paper introduces CSLib, a formal verification framework, which is a powerful tool for data science management and building robust, reliable systems. The use of Lean and formal verification aligns with a focus on rigorous data analysis and model validation.
Due to your Interest in Data Science Management
Coming from Amazon, this paper introduces CSLib, a formal verification framework, which is a powerful tool for data science management and building robust, reliable systems. The use of Lean and formal verification aligns with a focus on rigorous data analysis and model validation.
Hong Kong University of Science and Technology HKUST
AI Insights
- Data quality issues: LLMs can perpetuate biases and inaccuracies present in training data. (ML: 0.99)
- As the use of LLMs becomes more widespread, it is essential to address concerns around data quality, bias, and transparency. (ML: 0.99)
- The use of Large Language Models (LLMs) is becoming increasingly prevalent in data analysis and visualization. (ML: 0.98)
- Ontology Matching: The process of matching concepts from different ontologies (formal representations of knowledge) to establish relationships between them. (ML: 0.97)
- Several studies have demonstrated the effectiveness of LLM-based agents in automating tasks such as data visualization, chart generation, and question answering. (ML: 0.95)
- Large Language Models (LLMs): A type of artificial intelligence model that can process and understand human language to generate text or perform other tasks. (ML: 0.95)
- Researchers are exploring various applications of LLMs, including data cleaning, data standardization, ontology matching, query rewriting, and database knob tuning. (ML: 0.95)
- Data+ AI Ecosystems: An integrated system that combines data storage, processing, and analysis with artificial intelligence capabilities. (ML: 0.95)
- The integration of LLMs into data analysis and visualization is a rapidly evolving field with significant potential for innovation and improvement. (ML: 0.94)
- Further research is needed to fully understand the capabilities and limitations of LLM-based agents in various applications. (ML: 0.93)
Abstract
Data agents are an emerging paradigm that leverages large language models (LLMs) and tool-using agents to automate data management, preparation, and analysis tasks. However, the term "data agent" is currently used inconsistently, conflating simple query responsive assistants with aspirational fully autonomous "data scientists". This ambiguity blurs capability boundaries and accountability, making it difficult for users, system builders, and regulators to reason about what a "data agent" can and cannot do. In this tutorial, we propose the first hierarchical taxonomy of data agents from Level 0 (L0, no autonomy) to Level 5 (L5, full autonomy). Building on this taxonomy, we will introduce a lifecycleand level-driven view of data agents. We will (1) present the L0-L5 taxonomy and the key evolutionary leaps that separate simple assistants from truly autonomous data agents, (2) review representative L0-L2 systems across data management, preparation, and analysis, (3) highlight emerging Proto-L3 systems that strive to autonomously orchestrate end-to-end data workflows to tackle diverse and comprehensive data-related tasks under supervision, and (4) discuss forward-looking research challenges towards proactive (L4) and generative (L5) data agents. We aim to offer both a practical map of today's systems and a research roadmap for the next decade of data-agent development.
Data agents are an emerging paradigm that leverages large language models (LLMs) and tool-using agents to automate data management, preparation, and analysis tasks. However, the term "data agent" is currently used inconsistently, conflating simple query responsive assistants with aspirational fully autonomous "data scientists". This ambiguity blurs capability boundaries and accountability, making it difficult for users, system builders, and regulators to reason about what a "data agent" can and cannot do. In this tutorial, we propose the first hierarchical taxonomy of data agents from Level 0 (L0, no autonomy) to Level 5 (L5, full autonomy). Building on this taxonomy, we will introduce a lifecycleand level-driven view of data agents. We will (1) present the L0-L5 taxonomy and the key evolutionary leaps that separate simple assistants from truly autonomous data agents, (2) review representative L0-L2 systems across data management, preparation, and analysis, (3) highlight emerging Proto-L3 systems that strive to autonomously orchestrate end-to-end data workflows to tackle diverse and comprehensive data-related tasks under supervision, and (4) discuss forward-looking research challenges towards proactive (L4) and generative (L5) data agents. We aim to offer both a practical map of today's systems and a research roadmap for the next decade of data-agent development.
Why we are recommending this paper?
Due to your Interest in Direction on Data Science Organizations
Due to your Interest in Direction on Data Science Organizations
Palo Alto Networks
AI Insights
- While the RL agent learns from engagement signals, incorporating cognitive load and user satisfaction as explicit reward components could further improve adaptation quality. (ML: 0.96)
- The current framework assumes stable user behavior distributions and relies on anonymized interaction data, which may not capture all contextual factors. (ML: 0.96)
- Predictive modeling: A technique used to forecast user behavior based on past interactions and contextual factors. (ML: 0.95)
- Results indicate that AI-driven personalization increases engagement metrics by up to 30% while reducing average interaction latency. (ML: 0.95)
- Reinforcement learning: An AI approach that enables systems to learn from rewards or penalties received after taking actions in an environment. (ML: 0.94)
- A key insight from this work is the synergistic relationship between sequence prediction and reward-based optimization, which balances immediacy and sustained engagement. (ML: 0.90)
- The proposed system integrates predictive modeling and reinforcement learning within a unified personalization loop to enhance front-end adaptability. (ML: 0.90)
- The study presents a concrete step toward practical, real-time AI-driven personalization in production-scale front-end systems. (ML: 0.86)
- Unified personalization loop: A framework that combines predictive modeling, reinforcement learning, and adaptive rendering to continuously improve UI adaptation. (ML: 0.86)
- By coupling prediction, optimization, and adaptive rendering within a single feedback architecture, it establishes a foundation for next-generation user interfaces that evolve intelligently alongside their users. (ML: 0.82)
Abstract
Front-end personalization has traditionally relied on static designs or rule-based adaptations, which fail to fully capture user behavior patterns. This paper presents an AI driven approach for dynamic front-end personalization, where UI layouts, content, and features adapt in real-time based on predicted user behavior. We propose three strategies: dynamic layout adaptation using user path prediction, content prioritization through reinforcement learning, and a comparative analysis of AI-driven vs. rule-based personalization. Technical implementation details, algorithms, system architecture, and evaluation methods are provided to illustrate feasibility and performance gains.
Front-end personalization has traditionally relied on static designs or rule-based adaptations, which fail to fully capture user behavior patterns. This paper presents an AI driven approach for dynamic front-end personalization, where UI layouts, content, and features adapt in real-time based on predicted user behavior. We propose three strategies: dynamic layout adaptation using user path prediction, content prioritization through reinforcement learning, and a comparative analysis of AI-driven vs. rule-based personalization. Technical implementation details, algorithms, system architecture, and evaluation methods are provided to illustrate feasibility and performance gains.
Why we are recommending this paper?
Due to your Interest in Personalization
Due to your Interest in Personalization
Seoul National University
AI Insights
- Vision-Language Models (VLMs): A type of artificial intelligence model that combines natural language processing and computer vision capabilities. (ML: 0.97)
- The paper discusses the concept of contextualized visual personalization in vision-language models (VLMs), which involves tailoring the model's responses to a user's specific context and preferences. (ML: 0.97)
- MCQA Accuracy: The accuracy of a VLM's ability to answer multiple-choice questions accurately. (ML: 0.97)
- Contextualized Visual Personalization: The ability of a VLM to tailor its responses to a user's specific context and preferences. (ML: 0.96)
- The paper concludes by emphasizing the need for more research on contextualized visual personalization and the development of new evaluation metrics and benchmarks to assess VLMs' performance in this area. (ML: 0.96)
- The authors highlight the importance of addressing issues such as object hallucination, where the model generates objects that are not present in the input image, and entity name inclusion, where the model fails to include relevant entities in its response. (ML: 0.93)
- The authors propose a new benchmark for evaluating VLMs' ability to personalize their responses, which includes tasks such as object hallucination, entity name inclusion, and positive MCQA accuracy. (ML: 0.93)
- The paper also discusses various techniques for improving VLMs' personalization capabilities, including retrieval-augmented personalization, knowledge-augmented large language models, and multi-concept customization of text-to-image diffusion. (ML: 0.92)
- Entity Name Inclusion: When a VLM fails to include relevant entities in its response. (ML: 0.89)
- Object Hallucination: When a VLM generates objects that are not present in the input image. (ML: 0.83)
Abstract
Despite recent progress in vision-language models (VLMs), existing approaches often fail to generate personalized responses based on the user's specific experiences, as they lack the ability to associate visual inputs with a user's accumulated visual-textual context. We newly formalize this challenge as contextualized visual personalization, which requires the visual recognition and textual retrieval of personalized visual experiences by VLMs when interpreting new images. To address this issue, we propose CoViP, a unified framework that treats personalized image captioning as a core task for contextualized visual personalization and improves this capability through reinforcement-learning-based post-training and caption-augmented generation. We further introduce diagnostic evaluations that explicitly rule out textual shortcut solutions and verify whether VLMs truly leverage visual context. Extensive experiments demonstrate that existing open-source and proprietary VLMs exhibit substantial limitations, while CoViP not only improves personalized image captioning but also yields holistic gains across downstream personalization tasks. These results highlight CoViP as a crucial stage for enabling robust and generalizable contextualized visual personalization.
Despite recent progress in vision-language models (VLMs), existing approaches often fail to generate personalized responses based on the user's specific experiences, as they lack the ability to associate visual inputs with a user's accumulated visual-textual context. We newly formalize this challenge as contextualized visual personalization, which requires the visual recognition and textual retrieval of personalized visual experiences by VLMs when interpreting new images. To address this issue, we propose CoViP, a unified framework that treats personalized image captioning as a core task for contextualized visual personalization and improves this capability through reinforcement-learning-based post-training and caption-augmented generation. We further introduce diagnostic evaluations that explicitly rule out textual shortcut solutions and verify whether VLMs truly leverage visual context. Extensive experiments demonstrate that existing open-source and proprietary VLMs exhibit substantial limitations, while CoViP not only improves personalized image captioning but also yields holistic gains across downstream personalization tasks. These results highlight CoViP as a crucial stage for enabling robust and generalizable contextualized visual personalization.
Why we are recommending this paper?
Due to your Interest in Personalization
Due to your Interest in Personalization
Northwestern University
AI Insights
- Gradual resolution of uncertainty: the process by which consumers gradually learn about the quality and prices of products over time. (ML: 0.98)
- Benchmark model: a model without hidden fees used for comparison purposes. (ML: 0.95)
- The paper concludes that gradual resolution of uncertainty leads to higher prices and lower consumer welfare, but regulation can mitigate these effects. (ML: 0.95)
- It also highlights the importance of considering the impact of regulatory policies on equilibrium prices and consumer welfare in markets with hidden fees. (ML: 0.94)
- It finds that gradual resolution of uncertainty leads to higher prices and lower consumer welfare compared to a benchmark model without hidden fees. (ML: 0.94)
- However, regulation that prohibits hidden fees can lead to lower prices and higher consumer welfare. (ML: 0.93)
- The paper studies the effect of gradual resolution of uncertainty on prices and consumer welfare in a monopolistic competition model with hidden fees. (ML: 0.91)
- Hidden fees: fees charged by firms to consumers that are not explicitly disclosed. (ML: 0.91)
- Monopolistic competition: a market structure where firms compete with each other but also have some degree of market power. (ML: 0.88)
- The paper also examines the effect of regulation on equilibrium prices in unregulated and regulated markets and finds that it increases consumer surplus. (ML: 0.83)
Abstract
I introduce and study a nested search problem modeled as a tree structure that generalizes Weitzman (1979) in two ways: (1) search progresses incrementally, reflecting real-life scenarios where agents gradually acquire information about the prizes; and (2) the realization of prizes can be correlated, capturing similarities among them. I derive the optimal policy, which takes the form of an index solution. I apply this result to study monopolistic competition in a market with two stages of product inspection. My application illustrates that regulations on drip pricing lower equilibrium price and raise consumer surplus.
I introduce and study a nested search problem modeled as a tree structure that generalizes Weitzman (1979) in two ways: (1) search progresses incrementally, reflecting real-life scenarios where agents gradually acquire information about the prizes; and (2) the realization of prizes can be correlated, capturing similarities among them. I derive the optimal policy, which takes the form of an index solution. I apply this result to study monopolistic competition in a market with two stages of product inspection. My application illustrates that regulations on drip pricing lower equilibrium price and raise consumer surplus.
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.
- Bidding
- customer relationship management (crm) optimization
You can edit or add more interests any time.
💬 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