Murat Sensoy

AAMAS Conference 2026 Conference Paper

Enabling User Agency in Scalable Content Recommendations with Large Language Models

• Yucheng Li
• Gerrit J. J. Van den Burg
• Wei Liu
• Zhunxuan Wang
• Abhishek Tripathi
• Murat Sensoy

Existing content recommender systems usually depend on centrally stored interaction histories, creating vendor lock-in and disadvantagingnewerproviderswholacksufficientuserdata. Theyalsolimit users’ ability to understand, control, or edit how their preferences are represented, since profiles are learned as opaque latent vectors within provider-controlled models. We propose a user-centric alternative in which personal agents construct interpretable, editable preference profiles in natural language. Each profile item is associated with a learnable weight indicating its importance, and profiles are learned locally under full user control, laying the groundwork for high-quality personalization across multiple content providers. Recommendations are generated by matching contentwithweightedprofileembeddingsinasharedembeddingspace that is fine-tuned once using only content data and subsequently used by both content providers and personal agents. This design shifts profile ownership to users while maintaining the efficiency of existing recommender systems, as online recommendation reduces to approximate nearest-neighbor search. It further lowers the barrier for new providers, who only need to embed their content into the shared space — personalization naturally emerges from user-side profile embeddings optimized by personal agents to retrieve the most relevant content. Experiments on the MIND and Goodreads datasets show that our system outperforms strong baselines while providing transparency and editability — reimagining personalization as a process owned and controlled by the user.

AAMAS Conference 2026 Conference Paper

LLMEvalRec: An Agentic Framework for Simulating Users to Evaluate News Recommendation Systems

• Yao Ma
• Abhishek Tripathi
• Samuel Louvan
• Wei Liu
• Murat Sensoy

Evaluating news recommendation systems (NRS) presents unique challenges due to their dynamic and interactive nature coupled with evolving user interests. In the early stages of development, when user bases and historical data are scarce, it is difficult to conduct meaningful offline and online evaluations. This cold-start evaluation challenge hinders data-driven decision-making for product development and deployment. To address this, we propose LLMEvalRec, a framework that leverages Large Language Model (LLM) agents to simulate user behavior for NRS evaluation. Our approach features generative agents that automatically generate user profiles from a small number of user reading histories and perform realistic actions, while introducing the Guided Episodic Search (GUES) algorithm, which guides the automated prompt optimization process by exploring human prompt engineering practices. Experiments demonstrate that LLMEvalRec-generated data achieves 0. 97 Spearman correlation with real evaluation rankings, significantly outperforming baseline simulators (0. 4 and -0. 05), and successfully predicts relative performance trends across both MIND benchmark and real customer datasets. Production environment validation shows consistent alignment between simulated metrics and real click-through rate (CTR) improvements.

AIJ Journal 2023 Journal Article

DEED: DEep Evidential Doctor

• Awais Ashfaq
• Markus Lingman
• Murat Sensoy
• Sławomir Nowaczyk

As Deep Neural Networks (DNN) make their way into safety-critical decision processes, it becomes imperative to have robust and reliable uncertainty estimates for their predictions for both in-distribution and out-of-distribution (OOD) examples. This is particularly important in real-life high-risk settings such as healthcare, where OOD examples (e. g. , patients with previously unseen or rare labels, i. e. , diagnoses) are frequent, and an incorrect clinical decision might put human life in danger, in addition to having severe ethical and financial costs. While evidential uncertainty estimates for deep learning have been studied for multi-class problems, research in multi-label settings remains untapped. In this paper, we propose a DEep Evidential Doctor (DEED), which is a novel deterministic approach to estimate multi-label targets along with uncertainty. We achieve this by placing evidential priors over the original likelihood functions and directly estimating the parameters of the evidential distribution using a novel loss function. Additionally, we build a redundancy layer (particularly for high uncertainty and OOD examples) to minimize the risk associated with erroneous decisions based on dubious predictions. We achieve this by learning the mapping between the evidential space and a continuous semantic label embedding space via a recurrent decoder. Thereby inferring, even in the case of OOD examples, reasonably close predictions to avoid catastrophic consequences. We demonstrate the effectiveness of DEED on a digit classification task based on a modified multi-label MNIST dataset, and further evaluate it on a diagnosis prediction task from a real-life electronic health record dataset. We highlight that in terms of prediction scores, our approach is on par with the existing state-of-the-art having a clear advantage of generating reliable, memory and time-efficient uncertainty estimates with minimal changes to any multi-label DNN classifier.

AAAI Conference 2020 Conference Paper

Uncertainty-Aware Deep Classifiers Using Generative Models

• Murat Sensoy
• Lance Kaplan
• Federico Cerutti
• Maryam Saleki

Deep neural networks are often ignorant about what they do not know and overconfident when they make uninformed predictions. Some recent approaches quantify classification uncertainty directly by training the model to output high uncertainty for the data samples close to class boundaries or from the outside of the training distribution. These approaches use an auxiliary data set during training to represent out-ofdistribution samples. However, selection or creation of such an auxiliary data set is non-trivial, especially for high dimensional data such as images. In this work we develop a novel neural network model that is able to express both aleatoric and epistemic uncertainty to distinguish decision boundary and out-of-distribution regions of the feature space. To this end, variational autoencoders and generative adversarial networks are incorporated to automatically generate out-of-distribution exemplars for training. Through extensive analysis, we demonstrate that the proposed approach provides better estimates of uncertainty for in- and out-of-distribution samples, and adversarial examples on well-known data sets against state-ofthe-art approaches including recent Bayesian approaches for neural networks and anomaly detection methods.

NeurIPS Conference 2018 Conference Paper

Evidential Deep Learning to Quantify Classification Uncertainty

• Murat Sensoy
• Lance Kaplan
• Melih Kandemir

Deterministic neural nets have been shown to learn effective predictors on a wide range of machine learning problems. However, as the standard approach is to train the network to minimize a prediction loss, the resultant model remains ignorant to its prediction confidence. Orthogonally to Bayesian neural nets that indirectly infer prediction uncertainty through weight uncertainties, we propose explicit modeling of the same using the theory of subjective logic. By placing a Dirichlet distribution on the class probabilities, we treat predictions of a neural net as subjective opinions and learn the function that collects the evidence leading to these opinions by a deterministic neural net from data. The resultant predictor for a multi-class classification problem is another Dirichlet distribution whose parameters are set by the continuous output of a neural net. We provide a preliminary analysis on how the peculiarities of our new loss function drive improved uncertainty estimation. We observe that our method achieves unprecedented success on detection of out-of-distribution queries and endurance against adversarial perturbations.

AAMAS Conference 2016 Conference Paper

SOBE: Source Behavior Estimation for Subjective Opinions in Multiagent Systems (Extended Abstract)

• Murat Sensoy
• Lance Kaplan
• Geeth De Mel
• Taha D. Gunes

In cooperative or hostile environments, agents communicate their subjective opinions about various phenomenon. However, sources of these opinions may not always be competent and honest but more likely erroneous or even malicious. Furthermore, malicious sources may adopt certain behaviors to mislead the decision maker in a specific way. Fortunately, the reports of such misleading sources are correlated to ground truth. In this work, we propose to learn statistically meaningful opinion transformations that represent various behaviors of information sources. Then, we exploit these transformations while fusing opinions from unreliable sources. We show that our approach can be used to determine set of transformations that may lead to more accurate estimation of the truth.

EAAI Journal 2013 Journal Article

A hybrid reasoning mechanism for effective sensor selection for tasks

• Geeth De Mel
• Murat Sensoy
• Wamberto Vasconcelos
• Timothy J. Norman

In this paper, we present Ontological Logic Programming (OLP), a novel approach that combines logic programming with ontological reasoning. OLP enables the use of ontological terms (i. e. , individuals, classes and properties) directly within logic programmes. The interpretation of these terms is delegated to an ontology reasoner during the interpretation of the programme. Unlike similar approaches, OLP makes use of the full capacity of both ontological reasoning and logic programming. We evaluate the computational properties of OLP in different settings and show that its performance can be significantly improved using caching mechanisms. We then introduce a comprehensive sensor-task selection solution based on OLP and discuss the benefits one can obtain by using OLP. The solution is based on a set of interlinking ontologies that capture the crucial domain knowledge of sensor networks. We then make use of OLP to create and manage complex concepts in the domain as well as to implement effective resource-task assignment algorithms, which compute appropriate resources for tasks such that they sufficiently cover the tasks needs. We compare the advantages of OLP with a knowledge-based set-covering mechanism for resource-task selection.

IS Journal 2013 Journal Article

Agilely Assigning Sensing Assets to Mission Tasks in a Coalition Context

• Alun Preece
• Tim Norman
• Geeth De Mel
• Diego Pizzocaro
• Murat Sensoy
• Tien Pham

When managing intelligence, surveillance, and reconnaissance (ISR) operations in a coalition context, assigning available sensing assets to mission tasks can be challenging. The authors' approach to ISR asset assignment uses ontologies, allocation algorithms, and a service-oriented architecture.

AAAI Conference 2012 Conference Paper

Querying Linked Ontological Data through Distributed Summarization

• Achille Fokoue
• Felipe Meneguzzi
• Murat Sensoy
• Jeff Pan

As the semantic web expands, ontological data becomes distributed over a large network of data sources on the Web. Consequently, evaluating queries that aim to tap into this distributed semantic database necessitates the ability to consult multiple data sources efficiently. In this paper, we propose methods and heuristics to efficiently query distributed ontological data based on a series of properties of summarized data. In our approach, each source summarizes its data as another RDF graph, and relevant section of these summaries are merged and analyzed at query evaluation time. We show how the analysis of these summaries enables more efficient source selection, query pruning and transformation of expensive distributed joins into local joins.

AAMAS Conference 2010 Conference Paper

Distributed Semantic Search for the Web: A Multiagent Approach

• Murat Sensoy

In this paper, we propose to combine Semantic Web technologiesand multiagent systems in a novel way to enable users to locate andshare URLs relevant to their search interests.

AAMAS Conference 2010 Conference Paper

Flexible Task Resourcing for Intelligent Agents

• Murat Sensoy
• Wamberto W. Vasconcelos
• Timothy Norman

In many applications, tasks can be delegated to intelligent agents. In order to carry out a task, an agent should reason about what typesof resources the task requires. However, determining the right resource types requires extensive expertise and domain knowledge. In this paper, we propose means to automate the selection of resource types that are required to fulfill tasks. Our approach combines ontological reasoning and logic programming for a flexiblematchmaking of resources to tasks. Using the proposed approach, intelligent agents can autonomously reason about the resources andtasks in various real-life settings. Using a case-study, we describeand evaluate how agents can use the proposed approach to promoteresource sharing. Our evaluations show that the proposed approachis efficient and very useful for multi-agent systems.

AAMAS Conference 2008 Conference Paper

A Cooperation-Based Approach For Evolution Of Service Ontologies

• Murat Sensoy
• Pinar Yolum

Communication among agents requires a common vocabulary to facilitate successful information exchange. One way to achieve this is to assume the existence of a common ontology among communicating agents. However, this is a strong assumption, because agents may experience situations that result in independent evolution of their ontologies. When this is the case, agents need to form common grounds to enable communication. Accordingly, this paper proposes an approach in which agents can add new service concepts into their service ontologies and teach others services from their ontologies by exchanging service descriptions. This leads to a society of agents with different but overlapping ontologies where mutually accepted services emerge based on agents’ exchange of service descriptions. Our simulations of societies show that allowing cooperative evolution of local service ontologies facilitates better representation of agents’ needs. Further, through cooperation, not only more useful services emerge over time, but also ontologies of agents having similar service needs become aligned gradually.

ECAI Conference 2008 Conference Paper

Active Concept Learning For Ontology Evolution

• Murat Sensoy
• Pinar Yolum

This paper proposes an approach that enables agents to teach each other concepts from their ontologies using examples. Unlike other concept learning approaches, our approach enables the learner to elicit the most informative examples interactively from the teacher. Hence, the learner participates to the learning process actively. We empirically compare the proposed approach with the previous concept learning approaches. Our experiments show that using the proposed approach, agents can learn new concepts successfully and with fewer examples.

AAAI Conference 2007 Short Paper

A Framework for Ontology-Based Service Selection in Dynamic Environments

• Murat Sensoy

Previous approaches to service selection are mainly based on capturing and exchanging the ratings of consumers to providers. However, ratings reflect tastes of the raters. Therefore, service selection using ratings may mislead the consumers having a taste different than that of the raters. We propose to use experiences instead of the ratings. Experiences are the representation of what is requested by a consumer and what is received at the end. Unlike ratings, experiences do not reflect the opinion of the others, but the actual story between consumers and providers concerning a service demand. Using experiences, the consumer models the services of a provider for a specific service demand and selects the provider that is expected to satisfy the consumer the most. Our simulations show that proposed approach significantly increases the overall satisfaction of the service consumers.

AAMAS Conference 2006 Conference Paper

A Context-Aware Approach for Service Selection Using Ontologies

• Murat Sensoy
• Pinar Yolum