arXiv subjects

Maryam Saeedi

Publications and source records attributed to Maryam Saeedi.

4 recordsLinked to original sources

arXiv2025

Divide and Confer: Aggregating Information without Verification

We examine receiver-optimal mechanisms for aggregating information divided across many biased senders. Each sender privately observes an unconditionally independent signal about an unknown state, so no sender can verify another's report. A receiver makes a binary accept/reject decision that determines the players' payoffs via the state. When information is divided across a small population, and bias is low, the receiver-optimal mechanism coincides with the sender-preferred allocation, and can be implemented by letting senders confer privately before reporting. However, for larger populations, the receiver can benefit from the informational divide. We introduce a novel incentive-compatibility-in-the-large approach to solve the high-dimensional mechanism design problem for the large-population limit. Using this, we show that optimal mechanisms converge to one that depends only on the accept payoff and punishes excessive consensus in the direction of the common bias. These surplus burning punishments lead to payoffs that are bounded away from the first-best.

econ.TH↗

arXiv2024

Getting the Agent to Wait

We examine the strategic interaction between an expert (principal) maximizing engagement and an agent seeking swift information. Our analysis reveals: When priors align, relative patience determines optimal disclosure -- impatient agents induce gradual revelation, while impatient principals cause delayed, abrupt revelation. When priors disagree, catering to the bias often emerges, with the principal initially providing signals aligned with the agent's bias. With private agent beliefs, we observe two phases: one engaging both agents, followed by catering to one type. Comparing personalized and non-personalized strategies, we find faster information revelation in the non-personalized case, but higher quality information in the personalized case.

econ.TH↗

arXiv2020

Optimal Rating Design under Moral Hazard

We study optimal rating design under moral hazard and strategic manipulation. An intermediary observes a noisy indicator of effort and commits to a rating policy that shapes market beliefs and pay. Whether optimal ratings reveal or censor information depends on how effort shifts the outcome distribution: when effort increases tail risk, optimal ratings use lower censorship, pooling poor outcomes to encourage risk-taking; when effort reduces tail risk, upper censorship discourages negligence. In multi-task settings with window dressing, a monotone relative informativeness property again delivers upper- or lower-censorship ratings. In an application to redistributive test design, optimal tests can feature mid-censorship. These results follow from a general characterization of optimal ratings via concavification of a gain function, accommodating violations of monotone likelihood ratios and distributional concerns.

econ.TH↗

arXiv2020

A Novel and Reliable Deep Learning Web-Based Tool to Detect COVID-19 Infection from Chest CT-Scan

The corona virus is already spread around the world in many countries, and it has taken many lives. Furthermore, the world health organization (WHO) has announced that COVID-19 has reached the global epidemic stage. Early and reliable diagnosis using chest CT-scan can assist medical specialists in vital circumstances. In this work, we introduce a computer aided diagnosis (CAD) web service to detect COVID- 19 online. One of the largest public chest CT-scan databases, containing 746 participants was used in this experiment. A number of well-known deep neural network architectures consisting of ResNet, Inception and MobileNet were inspected to find the most efficient model for the hybrid system. A combination of the Densely connected convolutional network (DenseNet) in order to reduce image dimensions and Nu-SVM as an anti-overfitting bottleneck was chosen to distinguish between COVID-19 and healthy controls. The proposed methodology achieved 90.80% recall, 89.76% precision and 90.61% accuracy. The method also yields an AUC of 95.05%. Ultimately a flask web service is made public through ngrok using the trained models to provide a RESTful COVID-19 detector, which takes only 39 milliseconds to process one image. The source code is also available at https://github.com/KiLJ4EdeN/COVID_WEB. Based on the findings, it can be inferred that it is feasible to use the proposed technique as an automated tool for diagnosis of COVID-19.

eess.IV↗