The ASPIRE Research Group

The Audio SPeech and Information REtrieval (ASPIRE) research group develops algorithms that analyze, extract meaningful information, and that make predictions from audio, speech and signal data. This is currently accomplished by developing novel algorithms that leverage advanced probabilistic, machine learning, and deep learning concepts. The research group works on projects that remove unwanted background noise from speech, predict human-level assessment of speech quality and intelligibility, and projects that develop mechanisms for ensuring audio and speech privacy for consumer electronic devices. These efforts have resulted in presentations and papers at top-tier venues, such as the IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP); the IEEE/ACM Transactions on Audio, Speech, and Language Processing (TASLP); the IEEE International Workshop on Machine Learning and Signal Processing (MLSP); the International Conference on Latent Variable Analysis and Signal Separation (LVA/ICA); and the Journal of the Acoustical Society of America (JASA), to name a few.

Another goal for this group is to aspire the next generation of researchers to pursue careers in computer science and machine learning, especially for individuals from traditionally underrepresented groups. Throughout the year, our members participate in various outreach efforts that introduce young students to computer science and that provide them will necessary skills for success.

Latest News


Congratulations to Khandokar Md. Nayem! He had his first paper, which is entitled "Incorporating intra-spectral dependencies with a recurrent output layer for improved speech enhancement," accepted to MLSP. Preliminary results from this effort were also recently orally presented at the Midwest Music and Audio Day (MMAD 2019), which was held here at IU.


Congratulations to Xuan Dong! His paper on a classification-aided framework for non-intrusive speech quality assessment was accepted to WASPAA.


Congratulations to Zhuohuang Zhang! He will present a paper on the "Impact of amplification on speech enhancement algorithms using an objective evaluation metric" at the International Congress on Acoustics (ICA 2019) in Aachen, Germany.


I'm excited to present joint work with Xuan Dong on a classification-aided framework for non-intrusive speech quality assessment, at the Midwest Music and Audio Day (MMAD 2019) that is being held at IU. I also look forward to seeing Khandokar Nayem's presentation on intra-spectral dependencies. Prof. Williamson help organize this one-day workshop along with other SICE faculty.


Let's welcome Daniel Quintans, Muhammad Asghar, and Chitrank Gupta to the ASPIRE research group. These undergraduates will be working in our group for the summer through an NSF-funded research experience for undergraduates (REU) and through IU's Global Talent Attraction Program (GTAP). The will be working on data collection and developing machine learning algorithms.


Our group recently received an IU FRSP seed-funding grant, to fund preliminary work on the importance of phase to individuals with hearing impairments. We look forward to begin this much needed work!


I'm extremely excited to now be a part of SICE's Data Science program!


Congratulations to EJ Seong! She recently presented a poster at the Midwest Security Workshop (MSW 2019). The title of her poster is "Boxing Attackers In: Towards Tangible Defenses against Eavesdropping Attacks."


Our abstract on the "Impact of Amplification on Speech Enhancement Algorithms using an Objective Evaluation Metric" was accepted to the Internation Congress on Acoustics (ICA) 2019 conference! Look forward to writing the full paper version.


Very excited to be a Grant Thornton (GT) Scholar and to be collaborating with GT, SPEA, and Kelly as part of GT-IDEA #GT Scholar #GT-IDEA


Congratulations to our group member, Zhuohuang Zhang, who has his first ICASSP publication! The title of his paper is, "OBJECTIVE COMPARISON OF SPEECH ENHANCEMENT ALGORITHMS WITH HEARING LOSS SIMULATION."


Our joint paper on "Building a Common Voice Corpus for Laiholh (Hakha Chin)" was accepted to ComputEL-3. This is just the beginning for addressing an extremely important problem. [PDF]


The future is bright for STEM. There were so many wonderful and intelligent young women at the OurCS #HelloResearch workshop. I'm so glad that I co-led one of the projects. [OurCS]


Our paper on phase-aware denoising was accepted to MLSP, which will be held in Aalborg, Denmark! [PDF]


Congratulations to Xuan Dong on getting his 1st publication! His work on long-term SNR estimation was accepted to the LVA ICA conference which will be held in the UK!


I'm excited to announce that the National Science Foundation (NSF) has decided to fund our grant for the CISE CRII program! This grant provides ~$175,000 and will help fund graduate students and allow us to make progress towards our research goals. Thanks NSF!


Prof. Williamson received a NVIDIA GPU grant valued at ~$2,000. This grant provides two NVIDIA TITAN Xp GPUs that will be installed in our private server


Prof. Williamson gave a poster talk on complex masking at the Midwest Music and Audio Day (MMAD) at Northwestern University


Our paper on the "Impact of Phase Estimation on Single-Channel Speech Separation Based on Time-Frequency Masking" was accepted for publication in the Journal of the Acoustical Society of America (JASA)


Prof. Williamson gave a talk to IU's Data Science Club about work on "Separating Speech from Background Noise using a Deep Neural Network and a Complex Mask".


Our paper on "Time-Frequency Masking in the Complex Domain for Speech Dereverberation and Denoising" was accepted for publication in IEEE Trans. on Audio, Speech, and Lang. Proc. (TASLP)


Our paper on "SPEECH DEREVERBERATION AND DENOISING USING COMPLEX RATIO MASKS" was accepted for publication in IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP) 2017


Prof. Williamson gave a talk at IU's Intelligent & Interactive Systems (IIS) Talk Series today, about our recent work on "Separating Speech from Background Noise using a Deep Neural Network and a Complex Mask". [Video Link]