AIConfPaper
← Search

Dimitrios Dimitriadis

19 accepted papers

2025

Do Not Design, Learn: A Trainable Scoring Function for Uncertainty Estimation in Generative LLMs

Duygu Nur Yaldiz, Yavuz Faruk Bakman, Baturalp Buyukates, Chenyang Tao, Anil Ramakrishna, Dimitrios Dimitriadis +2
NAACL 2025findings

Uncertainty estimation (UE) of generative large language models (LLMs) is crucial for evaluating the reliability of generated sequences. A significant subset of UE methods utilize token probabilities to assess uncertainty, aggregating multiple token probabilities into a single UE score using a scori…

Cited by 1Source ↗PDF ↗Scholar ↗
2025

Learning to Specialize: Joint Gating-Expert Training for Adaptive MoEs in Decentralized Settings

Yehya Farhat, Hamza ElMokhtar Shili, Fangshuo Liao, Chen Dun, Mirian Del Carmen Hipolito Garcia, Guoqing Zheng +4
NeurIPS 2025poster

Mixture-of-Experts (MoEs) achieve scalability by dynamically activating subsets of their components. Yet, understanding how expertise emerges through joint training of gating mechanisms and experts remains incomplete, especially in scenarios without clear task partitions. Motivated by inference cost…

Cited by 0Source ↗Scholar ↗
2024

CroMo-Mixup: Augmenting Cross-Model Representations for Continual Self-Supervised Learning

Erum Mushtaq*, Duygu Nur Yaldiz, Yavuz Faruk Bakman, Jie Ding, Chenyang Tao, Dimitrios Dimitriadis +1
ECCV 2024poster

"Continual self-supervised learning (CSSL) learns a series of tasks sequentially on the unlabeled data. Two main challenges of continual learning are catastrophic forgetting and task confusion. While CSSL problem has been studied to address the catastrophic forgetting challenge, little work has been…

Cited by 2Source ↗PDF ↗code ↗Scholar ↗
2024

Invariant Aggregator for Defending against Federated Backdoor Attacks

Xiaoyang Wang, Dimitrios Dimitriadis, Sanmi Koyejo, Shruti Tople
AISTATS 2024poster

Federated learning enables training high-utility models across several clients without directly sharing their private data. As a downside, the federated setting makes the model vulnerable to various adversarial attacks in the presence of malicious clients. Despite the theoretical and empirical succe…

Cited by 5Source ↗PDF ↗Scholar ↗
2024

MARS: Meaning-Aware Response Scoring for Uncertainty Estimation in Generative LLMs

Yavuz Faruk Bakman, Duygu Nur Yaldiz, Baturalp Buyukates, Chenyang Tao, Dimitrios Dimitriadis, Salman Avestimehr
ACL 2024long

Generative Large Language Models (LLMs) are widely utilized for their excellence in various tasks. However, their tendency to produce inaccurate or misleading outputs poses a potential risk, particularly in high-stakes environments. Therefore, estimating the correctness of generative LLM outputs is…

Cited by 13Source ↗PDF ↗code ↗Scholar ↗
2023

Efficient and Light-Weight Federated Learning via Asynchronous Distributed Dropout

Chen Dun, Mirian Hipolito, Chris Jermaine, Dimitrios Dimitriadis, Anastasios Kyrillidis
AISTATS 2023poster

Asynchronous learning protocols have regained attention lately, especially in the Federated Learning (FL) setup, where slower clients can severely impede the learning process. Herein, we propose AsyncDrop, a novel asynchronous FL framework that utilizes dropout regularization to handle device hetero…

Cited by 31Source ↗PDF ↗Scholar ↗
2023

Local or Global: Selective Knowledge Assimilation for Federated Learning with Limited Labels

Yae Jee Cho, Gauri Joshi, Dimitrios Dimitriadis
ICCV 2023poster

Many existing FL methods assume clients with fully-labeled data, while in realistic settings, clients have limited labels due to the expensive and laborious process of labeling. Limited labeled local data of the clients often leads to their local model having poor generalization abilities to their l…

Cited by 10PDF ↗Scholar ↗
2022

Heterogeneous Ensemble Knowledge Transfer for Training Large Models in Federated Learning

Yae Jee Cho, Andre Manoel, Gauri Joshi, Robert Sim, Dimitrios Dimitriadis
IJCAI 2022poster

Federated learning (FL) enables edge-devices to collaboratively learn a model without disclosing their private data to a central aggregating server. Most existing FL algorithms require models of identical architecture to be deployed across the clients and server, making it infeasible to train large…

Cited by 159Source ↗PDF ↗Scholar ↗
2022

UserIdentifier: Implicit User Representations for Simple and Effective Personalized Sentiment Analysis

Fatemehsadat Mireshghallah, Vaishnavi Shrivastava, Milad Shokouhi, Taylor Berg-Kirkpatrick, Robert Sim, Dimitrios Dimitriadis
NAACL 2022long

Global models are typically trained to be as generalizable as possible. Invariance to the specific user is considered desirable since models are shared across multitudes of users. However, these models are often unable to produce personalized responses for individual users, based on their data. Cont…

Cited by 39Source ↗PDF ↗Scholar ↗
2021

Ensemble Combination between Different Time Segmentations

Jeremy Heng Meng Wong, Dimitrios Dimitriadis, Ken'ichi Kumatani, Yashesh Gaur, George Polovets, Partha Parthasarathy +3
ICASSP 2021accepted

Hypothesis-level combination between multiple models can often yield gains in speech recognition. However, all models in the ensemble are usually restricted to use the same audio segmentation times. This paper proposes to generalise hypothesis-level combination, allowing the use of different audio s…

Cited by 0Source ↗Scholar ↗
2020

A Memory Augmented Architecture for Continuous Speaker Identification in Meetings

Nikolaos Flemotomos, Dimitrios Dimitriadis
ICASSP 2020accepted

We introduce and analyze a novel approach to the problem of speaker identification in multi-party recorded meetings. Given a speech segment and a set of available candidate profiles, a data-driven approach is proposed learning the distance relations between them, aiming at identifying the correct sp…

Cited by 0Source ↗Scholar ↗
2020

Combining Acoustics, Content and Interaction Features to Find Hot Spots in Meetings

Dave Makhervaks, William Hinthorn, Dimitrios Dimitriadis, Andreas Stolcke
ICASSP 2020accepted

Involvement hot spots have been proposed as a useful concept for meeting analysis and studied off and on for over 15 years. These are regions of meetings that are marked by high participant involvement, as judged by human annotators. However, prior work was either not conducted in a formal machine l…

Cited by 0Source ↗Scholar ↗
2019

Low-latency Speaker-independent Continuous Speech Separation

Takuya Yoshioka, Zhuo Chen, Changliang Liu, Xiong Xiao, Hakan Erdogan, Dimitrios Dimitriadis
ICASSP 2019accepted

Speaker independent continuous speech separation (SI-CSS) is a task of converting a continuous audio stream, which may contain overlapping voices of unknown speakers, into a fixed number of continuous signals each of which contains no overlapping speech segment. A separated, or cleaned, version of e…

Cited by 0Source ↗Scholar ↗
2019

Single-channel Speech Extraction Using Speaker Inventory and Attention Network

Xiong Xiao, Zhuo Chen, Takuya Yoshioka, Hakan Erdogan, Changliang Liu, Dimitrios Dimitriadis +2
ICASSP 2019accepted

Neural network-based speech separation has received a surge of interest in recent years. Previously proposed methods either are speaker independent or extract a target speaker's voice by using his or her voice snippet. In applications such as home devices or office meeting transcriptions, a possible…

Cited by 76Source ↗Scholar ↗
2018

Improving End-of-Turn Detection in Spoken Dialogues by Detecting Speaker Intentions as a Secondary Task

Zakaria Aldeneh, Dimitrios Dimitriadis, Emily Mower Provost
ICASSP 2018accepted

This work focuses on the use of acoustic cues for modeling turn-taking in dyadic spoken dialogues. Previous work has shown that speaker intentions (e.g., asking a question, uttering a backchannel, etc.) can influence turn-taking behavior and are good predictors of turn-transitions in spoken dialogue…

Cited by 0Source ↗Scholar ↗
2017

Speaker diarization: A perspective on challenges and opportunities from theory to practice

Kenneth Church, Weizhong Zhu, Josef Vopicka, Jason W. Pelecanos, Dimitrios Dimitriadis, Petr Fousek
ICASSP 2017accepted

This paper discusses some challenges and opportunities in developing a speaker diarization system for operation on real world call center telephony data. We contrast some of the differences between a standard data set akin to NIST evaluations and those found in call centers. In exploring these diffe…

Cited by 9Source ↗Scholar ↗
2016

CNMF-based acoustic features for noise-robust ASR

Colin Vaz, Dimitrios Dimitriadis, Samuel Thomas, Shrikanth S. Narayanan
ICASSP 2016accepted

We present an algorithm using convolutive non-negative matrix factorization (CNMF) to create noise-robust features for automatic speech recognition (ASR). Typically in noise-robust ASR, CNMF is used to remove noise from noisy speech prior to feature extraction. However, we find that denoising introd…

Cited by 0Source ↗Scholar ↗
2016

On the importance of event detection for ASR

David Haws, Dimitrios Dimitriadis, George Saon, Samuel Thomas, Michael Picheny
ICASSP 2016accepted

The performance of modern large vocabulary continuous speech recognition (LVCSR) systems is heavily affected by segment boundaries, proper speaker identification of the segments, as well as removal of spurious data. We propose to use Long Short Term Memory (LSTM) recurrent neural networks to partiti…

Cited by 0Source ↗Scholar ↗