arXiv subjects

Reza Mirghaderi

Publications and source records attributed to Reza Mirghaderi.

3 recordsLinked to original sources

arXiv2026

An LLM-powered Agentic Recommendation System for Connected TV Content Discovery

Recommendation systems, from traditional multi-stage to recent unified generative architectures, face challenges in incorporating diverse contextual signals, such as trending topics, breaking news, cultural events, and cross-surface user activities, into their ranking pipelines. These systems are designed to consume structured behavioral signals with consistent schemas, and lack the reasoning capability to naturally process unstructured or heterogeneously formatted contextual information. Incorporating such signals typically requires feature engineering, bespoke data pipelines, and carefully tuned heuristics. In this paper, we present an LLM-powered agentic recommendation system designed for Connected TV (CTV) content discovery that addresses these limitations. Our system leverages the reasoning capabilities of large language models to naturally process and synthesize diverse signals across varying schemas and structures, eliminating much of the manual integration inherent in traditional ranking and retrieval systems. Recognizing that current LLM-based solutions still fall short of traditional machine learning models in several recommendation tasks, including retrieval efficiency, personalization precision, and scalability, we adopt an agentic architecture that orchestrates specialized components, allowing each sub-task to be handled by the most suitable method, whether LLM-based or traditional ML. The main contribution of this work is our engineering approach to successfully overcoming the practical limitations of enabling LLM for recommendation, particularly inference latency. We share insights from our work and discuss the trade-offs and lessons learned in building a hybrid system that combines the flexibility of LLMs with the performance of established recommendation techniques.

cs.IR↗

arXiv2024

Item-Language Model for Conversational Recommendation

Large-language Models (LLMs) have been extremely successful at tasks like complex dialogue understanding, reasoning and coding due to their emergent abilities. These emergent abilities have been extended with multi-modality to include image, audio, and video capabilities. Recommender systems, on the other hand, have been critical for information seeking and item discovery needs. Recently, there have been attempts to apply LLMs for recommendations. One difficulty of current attempts is that the underlying LLM is usually not trained on the recommender system data, which largely contains user interaction signals and is often not publicly available. Another difficulty is user interaction signals often have a different pattern from natural language text, and it is currently unclear if the LLM training setup can learn more non-trivial knowledge from interaction signals compared with traditional recommender system methods. Finally, it is difficult to train multiple LLMs for different use-cases, and to retain the original language and reasoning abilities when learning from recommender system data. To address these three limitations, we propose an Item-Language Model (ILM), which is composed of an item encoder to produce text-aligned item representations that encode user interaction signals, and a frozen LLM that can understand those item representations with preserved pretrained knowledge. We conduct extensive experiments which demonstrate both the importance of the language-alignment and of user interaction knowledge in the item encoder.

cs.IR↗

arXiv2012

Achievable Error Exponents in the Gaussian Channel with Rate-Limited Feedback

We investigate the achievable error probability in communication over an AWGN discrete time memoryless channel with noiseless delay-less rate-limited feedback. For the case where the feedback rate R_FB is lower than the data rate R transmitted over the forward channel, we show that the decay of the probability of error is at most exponential in blocklength, and obtain an upper bound for increase in the error exponent due to feedback. Furthermore, we show that the use of feedback in this case results in an error exponent that is at least RF B higher than the error exponent in the absence of feedback. For the case where the feedback rate exceeds the forward rate (R_FB \geq R), we propose a simple iterative scheme that achieves a probability of error that decays doubly exponentially with the codeword blocklength n. More generally, for some positive integer L, we show that a L-th order exponential error decay is achievable if R_FB \geq (L-1)R. We prove that the above results hold whether the feedback constraint is expressed in terms of the average feedback rate or per channel use feedback rate. Our results show that the error exponent as a function of R_FB has a strong discontinuity at R, where it jumps from a finite value to infinity.

cs.IT↗