Publications de Yunus Inan | EPFL Graph Search

Causal Influence in Federated Edge Inference

In this paper, we consider a setting where heterogeneous agents with connectivity are performing inference using unlabeled streaming data. Observed data are only partially informative about the target variable of interest. In order to overcome the uncertai ...

IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC2024

On the Arithmetic and Geometric Fusion of Beliefs for Distributed Inference

Ali H. Sayed, Emre Telatar, Mert Kayaalp, Yunus Inan

We study the asymptotic learning rates of belief vectors in a distributed hypothesis testing problem under linear and log-linear combination rules. We show that under both combination strategies, agents are able to learn the truth exponentially fast, with ...

Ieee-Inst Electrical Electronics Engineers Inc2024

Causal Impact Analysis for Asynchronous Decision Making

Ali H. Sayed, Mert Kayaalp, Yunus Inan

We consider a collaborative decision-making framework where heterogeneous agents receive streaming and partially informative observations. We consider two asynchronous scenarios that differ based on the agents' participation patterns and the fusion center' ...

IEEE2024

Time vs. Truth: Age-Distortion Tradeoffs and Strategies for Distributed Inference

Yunus Inan

In 1948, Claude Shannon laid the foundations of information theory, which grew out of a study to find the ultimate limits of source compression, and of reliable communication. Since then, information theory has proved itself not only as a quest to find the ...

EPFL2023

Age-Optimal Causal Labeling of Memoryless Processes

Emre Telatar, Yunus Inan

We consider a problem of labeling a memoryless temporal point process. The labelings have to be done causally. We study the tradeoff between the rate and age of the labeled process and characterize the optimal tradeoff curve. We show that the optimal curve ...

IEEE2022

A Fundamental Limit of Distributed Hypothesis Testing Under Memoryless Quantization

Ali H. Sayed, Emre Telatar, Mert Kayaalp, Yunus Inan

We consider a distributed binary hypothesis testing setup where multiple nodes send quantized information to a central processor, which is oblivious to the nodes' statistics. We study the regime where the missed detection (type-II error) probability decays ...

IEEE2022

Optimal Policies for Age and Distortion in a Discrete-Time Model

Emre Telatar, Yunus Inan, Reka Inovan

We propose a simple model to study the tradeoff between timeliness and distortion, where different pieces of data have a different cost of not being sent. We pose the question of finding the optimal tradeoff as a policy design problem amenable to dynamic p ...

IEEE2021

Finite-Level Quantization Procedures for Construction and Decoding of Polar Codes

Emre Telatar, Yunus Inan

We consider finite-level, symmetric quantization procedures for construction and decoding of polar codes. Whether polarization occurs in the presence of quantization is not known in general. In [1], it is shown that a simple three-level quantization proced ...

IEEE2020