arXiv subjects

Negin Golrezaei

Publications and source records attributed to Negin Golrezaei.

At least 19 recordsLinked to original sources

arXiv2026

MACS: A Hybrid Multi-Agent Framework for Reliable Conversational E-Commerce Recommendation

Conversational recommendation for e-commerce is increasingly mediated by large language models (LLMs), yet many real-world deployments operate under a stricter requirement: recommendations must be drawn only from a merchant's fixed catalog, without web search or unsupported product claims. In this setting, the main challenge is reliability under hard constraints: the system must satisfy user requirements, remain grounded in available inventory, and preserve preferences across multiple conversational turns. We present MACS (Multi-Agent Commerce System), a hybrid multi-agent framework for reliable conversational recommendation in fixed-catalog settings. MACS uses LLMs for language-facing tasks such as interpreting user requests, eliciting preferences, and generating responses, while correctness-critical operations, including product retrieval, hard-constraint filtering, brand exclusion, and progressive relaxation, are executed deterministically by the merchant agent. A session-persistent preference layer tracks constraints across turns, enabling consistent handling of budget overwrites and exclusion reversals. On a 140-query single-turn benchmark, MACS achieves the highest pass rate (87.1%) and perfect brand compliance (1.000). On a 10-scenario multi-turn benchmark, MACS achieves the strongest macro Pass@5 (72% vs. 56% GPT+Catalog / 52% Gemini+Catalog) with zero constraint drift. The advantage is sharpest on exclusion reversal (100% vs. 20% / 0%) and constraint accumulation (100% vs. 60% / 40%). Mean judged response quality is similar across systems (0.751 vs. 0.736). These results suggest that hybrid architectures combining deterministic constraint enforcement with session-persistent preference tracking provide stronger reliability-oriented performance than catalog-bound prompt-only baselines in the fixed-catalog merchant setting.

cs.IR↗

arXiv2026

Market Design for AI: Beyond the Copyright Binary

How can we design a market of human-generated content for use in training AI models that both enables technological progress and preserves individual incentives for high-quality content creation? Existing approaches take polar positions: a "free-for-all" model based on fair use and a "strong intellectual property rights" model. We show that both fail: Free-for-all does not compensate creators, and---by modeling as a static Stackelberg game---strong intellectual property rights also underpower creative incentives. We find this especially true for more innovative creators, a phenomenon we term the "originality penalty." Extending this insight to a dynamic model, we find another market failure undermining AI model performance, even for an initially good model: Such a model induces greater reliance by humans on AI-assisted creation, resulting in homogenized content feeding back into training, which degrades the model performance---a "curse of precision." We further propose a market design with a data intermediary negotiating collectively with the AI firm and subsidizing innovative contributions, thus restoring efficiency.

econ.TH↗

arXiv2026

Efficiency, Feasibility, and Incentive-Awareness in Constrained Online Resource Allocation

We study the dynamic allocation of indivisible resources to strategic agents under long-term constraints, where the planner aims to maximize social welfare, satisfy multiple constraints, and elicit near-truthful reports. We find standard primal-dual methods fragile in this setting: agents easily manipulate their reports to distort dual variables, sacrificing social efficiency for individual utility. To address this, we propose the Incentive-Aware Primal-Dual (IAPD) framework. On the primal side, we integrate three components to suppress manipulation: a VCG-based payment neutralizes immediate misreporting benefits, while epoch-based lazy updates and random exploration together ensure potential future gains are outweighed by immediate penalties. On the dual side, to overcome a learning barrier due to lazy updates -- which we call the "price of incentives" -- we design a novel optimistic online learning algorithm, O-FTRL-FP. It utilizes a fixed-point oracle to resolve the circular dependency between optimistic dual variables and the resulting allocations. Ultimately, our mechanism attains $\tilde{\mathcal O}(\sqrt T)$ social welfare regret, satisfies all long-term constraints, and induces a near-truthful equilibrium. It also smoothly generalizes to multi-unit multi-demand allocation problems. Notably, this $\tilde{\mathcal O}(\sqrt T)$ regret near-matches the non-strategic $Ω(\sqrt T)$ lower bound, demonstrating that incentive-awareness can be accommodated at nearly no cost.

cs.GT↗

arXiv2026

Learning to Bid in Discriminatory Auctions with Budget Constraints

We study repeated bidding in multi-unit discriminatory (pay-as-bid) auctions for a single bidder with per-round utility equal to value minus $α$ times payment, where $α\in[0,1]$ is a cost-of-capital parameter. The bidder aims to maximize cumulative utility over $T$ rounds subject to a total budget $B$. The problem is challenging even without budgets: the action space is exponential in $M$, the maximum demand of the bidder and the valuation vector (context) varies over time. Exploiting a decomposition of utility across units, we develop polynomial-time learning algorithms based on shortest paths in a directed acyclic graph, obtaining sublinear regret under both full-information and bandit feedback. In the bandit setting, the regret is independent of the number of contexts due to complete cross-learning: observing the utility of the chosen action under the realized context reveals the utility for the same action under all counterfactual contexts. With budget constraints, when the average normalized per-round budget $ρ=\frac{B}{MT}<1$, we design a coupled primal-dual algorithm in which the DAG-based procedure uses dual-adjusted edge weights for primal updates, while online gradient descent updates the dual variable, yielding $ρ$-approximate sublinear regret. Finally, we give implementations whose per-round time and space are independent of the number of contexts, enabling scalability to large or even infinite context spaces.

cs.LG↗

arXiv2026

Policy Regret for Embedding Model Routing: Contextual Bandits with Low-Rank Experts

Modern recommendation systems increasingly rely on dynamically routing diverse queries to multiple embedding models. Despite its practical significance, this problem remains poorly understood under realistic conditions like adversarial queries, bandit feedback, and limited observability of models. We formalize embedding model routing as an adversarial contextual linear bandit with low-rank experts, where contexts are queries, actions are items, and experts are the embedding models working on low-rank latent representation spaces. We first establish that standard regret notions suffer from structural misspecification or statistical intractability, and we identify a log-quadratic policy class that is expressive enough to capture query-dependent model routing, yet structured enough to allow efficient online learning. Second, we propose a policy gradient algorithm called Hypentropy Policy Gradient (HPG). It provably adapts to the unknown low-rank structure under incomplete information and attains $\tilde{\mathcal O}(s\sqrt{M T})$ linearized policy regret -- where $s, M$, and $T$ are the intrinsic rank of the experts, the number of models, and the number of rounds -- thus avoiding a curse of dimensionality. Finally, we also provide an computationally efficient and parameter-free implementation of HPG.

cs.LG↗

arXiv2026

Online Resource Allocation with Convex-set Machine-Learned Advice

Decision-makers often have access to machine-learned predictions about future demand that can help guide online resource allocation decisions. However, such predictions may be inaccurate. We develop a framework for online resource allocation with potentially unreliable machine-learned advice, where the advice is represented as a convex uncertainty set for the demand vector rather than a single point estimate. We introduce a parameterized class of Pareto-optimal online algorithms that balance consistency and robustness. The consistent ratio measures performance when the advice is accurate, while the robust ratio measures performance under adversarial demand when the advice is inaccurate. For a target consistency level C, our algorithms maximize robustness subject to achieving at least consistency level C. Our approach extends classical protection-level algorithms by introducing adaptive protection levels that dynamically respond to uncertainty in the advice. We also provide a method for computing the maximum achievable consistency level. Numerical experiments demonstrate that our algorithms outperform benchmark methods, including approaches based solely on point forecasts, by effectively balancing worst-case and average-case performance.

cs.DS↗

arXiv2026

Optimal Contest Beyond Convexity

In the contest design problem, there are $n$ strategic contestants, each of whom decides an effort level. A contest designer with a fixed budget must then design a mechanism that allocates a prize $p_i$ to the $i$-th rank based on the outcome, to incentivize contestants to exert higher costly efforts and induce high-quality outcomes. In this paper, we significantly deepen our understanding of optimal mechanisms under general settings by considering nonconvex objectives in contestants' qualities. Notably, our results accommodate the following objectives: (i) any convex combination of user welfare (motivated by recommender systems) and the average quality of contestants, and (ii) arbitrary posynomials over quality, both of which may neither be convex nor concave. In particular, these subsume classic measures such as social welfare, order statistics, and (inverse) S-shaped functions, which have received little or no attention in the contest literature to the best of our knowledge. Surprisingly, across all these regimes, we show that the optimal mechanism is highly structured: it allocates potentially higher prize to the first-ranked contestant, zero to the last-ranked one, and equal prizes to the all intermediate contestants, i.e., $p_1 \ge p_2 = \ldots = p_{n-1} \ge p_n = 0$. Thanks to the structural characterization, we obtain a fully polynomial-time approximation scheme given a value oracle. Our technical results rely on Schur-convexity of Bernstein basis polynomial-weighted functions, total positivity and variation diminishing property. En route to our results, we obtain a surprising reduction from a structured high-dimensional nonconvex optimization to a single-dimensional optimization by connecting the shape of the gradient sequences of the objective function to the number of transition points in optimum, which might be of independent interest.

cs.GT↗

arXiv2026

Entropy Guided Diversification and Preference Elicitation in Agentic Recommendation Systems

Users on e-commerce platforms can be uncertain about their preferences early in their search. Queries to recommendation systems are frequently ambiguous, incomplete, or weakly specified. Agentic systems are expected to proactively reason, ask clarifying questions, and act on the user's behalf, which makes handling such ambiguity increasingly important. In existing platforms, ambiguity led to excessive interactions and question fatigue or overconfident recommendations prematurely collapsing the search space. We present an Interactive Decision Support System (IDSS) that addresses ambiguous user queries using entropy as a unifying signal. IDSS maintains a dynamically filtered candidate product set and quantifies uncertainty over item attributes using entropy. This uncertainty guides adaptive preference elicitation by selecting follow-up questions that maximize expected information gain. When preferences remain incomplete, IDSS explicitly incorporates residual uncertainty into downstream recommendations through uncertainty-aware ranking and entropy-based diversification, rather than forcing premature resolution. We evaluate IDSS using review-driven simulated users grounded in real user reviews, enabling a controlled study of diverse shopping behaviors. Our evaluation measures both interaction efficiency and recommendation quality. Results show that entropy-guided elicitation reduces unnecessary follow-up questions, while uncertainty-aware ranking and presentation yield more informative, diverse, and transparent recommendation sets under ambiguous intent. These findings demonstrate that entropy-guided reasoning provides an effective foundation for agentic recommendation systems operating under uncertainty.

cs.AI↗

arXiv2025

Learning Safe Strategies for Value Maximizing Buyers in Uniform Price Auctions

We study the bidding problem in repeated uniform price multi-unit auctions from the perspective of a value-maximizing buyer. The buyer aims to maximize their cumulative value over $T$ rounds while adhering to per-round return-on-investment (RoI) constraints in a strategic (or adversarial) environment. Using an $m$-uniform bidding format, the buyer submits $m$ bid-quantity pairs $(b_i, q_i)$ to demand $q_i$ units at bid $b_i$, with $m \ll M$ in practice, where $M$ denotes the maximum demand of the buyer. We introduce the notion of safe bidding strategies as those that satisfy the RoI constraints irrespective of competing bids. Despite the stringent requirement, we show that these strategies satisfy a mild no-overbidding condition, depend only on the valuation curve of the bidder, and the bidder can focus on a finite subset without loss of generality. Though the subset size is $O(M^m)$, we design a polynomial-time learning algorithm that achieves sublinear regret, both in full-information and bandit settings, relative to the hindsight-optimal safe strategy. We assess the robustness of safe strategies against the hindsight-optimal strategy from a richer class. We define the richness ratio $α\in (0,1]$ as the minimum ratio of the value of the optimal safe strategy to that of the optimal strategy from richer class and construct hard instances showing the tightness of $α$. Our algorithm achieves $α$-approximate sublinear regret against these stronger benchmarks. Simulations on semi-synthetic auction data show that empirical richness ratios significantly outperform the theoretical worst-case bounds. The proposed safe strategies and learning algorithm extend naturally to more nuanced buyer and competitor models.

cs.DS↗

arXiv2025

Fair Assortment Planning

Many online platforms, ranging from online retail stores to social media platforms, employ algorithms to optimize their offered assortment of items (e.g., products and contents). These algorithms often focus exclusively on achieving the platforms' objectives, highlighting items with the highest popularity or revenue. This approach, however, can compromise the equality of opportunities for the rest of the items, in turn leading to less content diversity and increased regulatory scrutiny for the platform. Motivated by this, we introduce and study a fair assortment planning problem that enforces equality of opportunities via pairwise fairness, which requires any two items to be offered similar outcomes. We show that the problem can be formulated as a linear program (LP), called (FAIR), that optimizes over the distribution of all feasible assortments. To find a near-optimal solution to (FAIR), we propose a framework based on the Ellipsoid method, which requires a polynomial-time separation oracle to the dual of the LP. We show that finding an optimal separation oracle to the dual problem is an NP-complete problem, and hence we propose a series of approximate separation oracles, which then result in a 1/2-approx. algorithm and an FPTAS for Problem (FAIR). The approximate separation oracles are designed by (i) showing the separation oracle to the dual of the LP is equivalent to solving an infinite series of parameterized knapsack problems, and (ii) leveraging the structure of knapsack problems. Finally, we perform numerical studies on both synthetic data and real-world MovieLens data, showcasing the effectiveness of our algorithms and providing insights into the platform's price of fairness.

cs.DS↗

arXiv2024

Learning in Repeated Multi-Unit Pay-As-Bid Auctions

Motivated by Carbon Emissions Trading Schemes, Treasury Auctions, Procurement Auctions, and Wholesale Electricity Markets, which all involve the auctioning of homogeneous multiple units, we consider the problem of learning how to bid in repeated multi-unit pay-as-bid auctions. In each of these auctions, a large number of (identical) items are to be allocated to the largest submitted bids, where the price of each of the winning bids is equal to the bid itself. In this work, we study the problem of optimizing bidding strategies from the perspective of a single bidder. Effective bidding in pay-as-bid (PAB) auctions is complex due to the combinatorial nature of the action space. We show that a utility decoupling trick enables a polynomial time algorithm to solve the offline problem where competing bids are known in advance. Leveraging this structure, we design efficient algorithms for the online problem under both full information and bandit feedback settings that achieve an upper bound on regret of $O(M \sqrt{T \log T})$ and $O(M T^{\frac{2}{3}} \sqrt{\log T})$ respectively, where $M$ is the number of units demanded by the bidder and $T$ is the total number of auctions. We accompany these results with a regret lower bound of $Ω(M\sqrt{T})$ for the full information setting and $Ω(M^{2/3}T^{2/3})$ for the bandit setting. We also present additional findings on the characterization of PAB equilibria. While the Nash equilibria of PAB auctions possess nice properties such as winning bid uniformity and high welfare \& revenue, they are not guaranteed under no regret learning dynamics. Nevertheless, our simulations suggest these properties hold anyways, regardless of Nash equilibrium existence. Compared to its uniform price counterpart, the PAB dynamics converge faster and achieve higher revenue, making PAB appealing whenever revenue holds significant social value.

cs.GT↗

arXiv2024

Interpolating Item and User Fairness in Multi-Sided Recommendations

Today's online platforms heavily lean on algorithmic recommendations for bolstering user engagement and driving revenue. However, these recommendations can impact multiple stakeholders simultaneously -- the platform, items (sellers), and users (customers) -- each with their unique objectives, making it difficult to find the right middle ground that accommodates all stakeholders. To address this, we introduce a novel fair recommendation framework, Problem (FAIR), that flexibly balances multi-stakeholder interests via a constrained optimization formulation. We next explore Problem (FAIR) in a dynamic online setting where data uncertainty further adds complexity, and propose a low-regret algorithm FORM that concurrently performs real-time learning and fair recommendations, two tasks that are often at odds. Via both theoretical analysis and a numerical case study on real-world data, we demonstrate the efficacy of our framework and method in maintaining platform revenue while ensuring desired levels of fairness for both items and users.

cs.IR↗

arXiv2024

Active Learning for Non-Parametric Choice Models

We study the problem of actively learning a non-parametric choice model based on consumers' decisions. We present a negative result showing that such choice models may not be identifiable. To overcome the identifiability problem, we introduce a directed acyclic graph (DAG) representation of the choice model. This representation provably encodes all the information about the choice model which can be inferred from the available data, in the sense that it permits computing all choice probabilities. We establish that given exact choice probabilities for a collection of item sets, one can reconstruct the DAG. However, attempting to extend this methodology to estimate the DAG from noisy choice frequency data obtained during an active learning process leads to inaccuracies. To address this challenge, we present an inclusion-exclusion approach that effectively manages error propagation across DAG levels, leading to a more accurate estimate of the DAG. Utilizing this technique, our algorithm estimates the DAG representation of an underlying non-parametric choice model. The algorithm operates efficiently (in polynomial time) when the set of frequent rankings is drawn uniformly at random. It learns the distribution over the most popular items among frequent preference types by actively and repeatedly offering assortments of items and observing the chosen item. We demonstrate that our algorithm more effectively recovers a set of frequent preferences on both synthetic and publicly available datasets on consumers' preferences, compared to corresponding non-active learning estimation algorithms. These findings underscore the value of our algorithm and the broader applicability of active-learning approaches in modeling consumer behavior.

cs.LG↗

arXiv2024

Learning and Collusion in Multi-unit Auctions

We consider repeated multi-unit auctions with uniform pricing, which are widely used in practice for allocating goods such as carbon licenses. In each round, $K$ identical units of a good are sold to a group of buyers that have valuations with diminishing marginal returns. The buyers submit bids for the units, and then a price $p$ is set per unit so that all the units are sold. We consider two variants of the auction, where the price is set to the $K$-th highest bid and $(K+1)$-st highest bid, respectively. We analyze the properties of this auction in both the offline and online settings. In the offline setting, we consider the problem that one player $i$ is facing: given access to a data set that contains the bids submitted by competitors in past auctions, find a bid vector that maximizes player $i$'s cumulative utility on the data set. We design a polynomial time algorithm for this problem, by showing it is equivalent to finding a maximum-weight path on a carefully constructed directed acyclic graph. In the online setting, the players run learning algorithms to update their bids as they participate in the auction over time. Based on our offline algorithm, we design efficient online learning algorithms for bidding. The algorithms have sublinear regret, under both full information and bandit feedback structures. We complement our online learning algorithms with regret lower bounds. Finally, we analyze the quality of the equilibria in the worst case through the lens of the core solution concept in the game among the bidders. We show that the $(K+1)$-st price format is susceptible to collusion among the bidders; meanwhile, the $K$-th price format does not have this issue.

cs.GT↗

arXiv2023

Non-Stationary Bandits with Auto-Regressive Temporal Dependency

Traditional multi-armed bandit (MAB) frameworks, predominantly examined under stochastic or adversarial settings, often overlook the temporal dynamics inherent in many real-world applications such as recommendation systems and online advertising. This paper introduces a novel non-stationary MAB framework that captures the temporal structure of these real-world dynamics through an auto-regressive (AR) reward structure. We propose an algorithm that integrates two key mechanisms: (i) an alternation mechanism adept at leveraging temporal dependencies to dynamically balance exploration and exploitation, and (ii) a restarting mechanism designed to discard out-of-date information. Our algorithm achieves a regret upper bound that nearly matches the lower bound, with regret measured against a robust dynamic benchmark. Finally, via a real-world case study on tourism demand prediction, we demonstrate both the efficacy of our algorithm and the broader applicability of our techniques to more complex, rapidly evolving time series.

cs.LG↗

arXiv2023

Upfront Commitment in Online Resource Allocation with Patient Customers

In many on-demand online platforms such as ride-sharing, grocery delivery, or shipping, some arriving agents are patient and willing to wait a short amount of time for the resource or service as long as there is an upfront guarantee that service will be ultimately provided within a certain delay. Motivated by this, we present a setting with patient and impatient agents who seek a resource or service that replenishes periodically. Impatient agents demand the resource immediately upon arrival while patient agents are willing to wait a short period conditioned on an upfront commitment to receive the resource. We study this setting under adversarial arrival models using a relaxed notion of competitive ratio. We present a class of POLYtope-based Resource Allocation (POLYRA) algorithms that achieve optimal or near-optimal competitive ratios. Such POLYRA algorithms work by consulting a particular polytope and only making decisions that guarantee the algorithm's state remains feasible in this polytope. When the number of agent types is either two or three, POLYRA algorithms can obtain the optimal competitive ratio. To design these polytopes, we construct an upper bound on the competitive ratio of any algorithm, which is characterized via a linear program (LP) that considers a collection of overlapping worst-case input sequences. Our designed POLYRA algorithms then mimic the optimal solution of this upper bound LP via its polytope's definition, obtaining the optimal competitive ratio. When there are more than three types, our overlapping worst-case input sequences do not necessarily result in an attainable competitive ratio, and so we present a class of simple and interpretable POLYRA algorithm which achieves at least 80% of the optimal competitive ratio. We complement our theoretical studies with numerical analysis which shows the efficiency of our algorithms beyond adversarial arrivals

cs.DS↗

arXiv2023

Optimal Learning for Structured Bandits

We study structured multi-armed bandits, which is the problem of online decision-making under uncertainty in the presence of structural information. In this problem, the decision-maker needs to discover the best course of action despite observing only uncertain rewards over time. The decision-maker is aware of certain convex structural information regarding the reward distributions; that is, the decision-maker knows the reward distributions of the arms belong to a convex compact set. In the presence such structural information, they then would like to minimize their regret by exploiting this information, where the regret is its performance difference against a benchmark policy that knows the best action ahead of time. In the absence of structural information, the classical upper confidence bound (UCB) and Thomson sampling algorithms are well known to suffer minimal regret. As recently pointed out, neither algorithms are, however, capable of exploiting structural information that is commonly available in practice. We propose a novel learning algorithm that we call "DUSA" whose regret matches the information-theoretic regret lower bound up to a constant factor and can handle a wide range of structural information. Our algorithm DUSA solves a dual counterpart of the regret lower bound at the empirical reward distribution and follows its suggested play. We show that this idea leads to the first computationally viable learning policy with asymptotic minimal regret for various structural information, including well-known structured bandits such as linear, Lipschitz, and convex bandits, and novel structured bandits that have not been studied in the literature due to the lack of a unified and flexible framework.

cs.LG↗

arXiv2023

Individual Welfare Guarantees in the Autobidding World with Machine-learned Advice

Online advertising channels have commonly focused on maximizing total advertiser value (or welfare) to enhance long-run retention and channel healthiness. Previous literature has studied auction design by incorporating machine learning predictions on advertiser values (also known as machine-learned advice) through various forms to improve total welfare. Yet, such improvements could come at the cost of individual bidders' welfare and do not shed light on how particular advertiser bidding strategies impact welfare. Motivated by this, we present an analysis on an individual bidder's welfare loss in the autobidding world for auctions with and without machine-learned advice, and also uncover how advertiser strategies relate to such losses. In particular, we demonstrate how ad platforms can utilize ML advice to improve welfare guarantee on the aggregate and individual bidder level by setting ML advice as personalized reserve prices when the platform consists of autobidders who maximize value while respecting a return-on-ad spent (ROAS) constraint. Under parallel VCG auctions with such ML advice-based reserves, we present a worst-case welfare lower-bound guarantee for an individual autobidder, and show that the lower-bound guarantee is positively correlated with ML advice quality as well the scale of bids induced by the autobidder's bidding strategies. Further, we prove an impossibility result showing that no truthful, and possibly randomized mechanism with anonymous allocations can achieve universally better individual welfare guarantees than VCG, in presence of personalized reserves based on ML-advice of equal quality. Moreover, we extend our individual welfare guarantee results to generalized first price (GFP) and generalized second price (GSP) auctions.

cs.GT↗