Nikita Singh
The 9th edition of the Global Testing Retreat 2024!
About Speaker
Currently, I thrive as a Quality Assurance professional with over 3 years of experience. Working across multiple retail domains, I ensure the quality delivery of products through manual and learning automated testing methodologies.
Beyond my technical pursuits, I find solace and passion in writing and poetry. Several of my poems have been published in anthologies, and I’ve even ventured into podcasting on Spotify.I’m currently channeling my creative energy into crafting my first solo book, a venture that excites and challenges me.
Additionally, I find joy and fulfillment in participating in marathons, pushing my physical and mental limits.In my professional journey, I also had the opportunity to work with the dental domain, managing client accounts from the UK and Canada.
This experience has enriched my understanding of global perspectives and sharpened my communication skills across diverse cultural backgrounds.My journey embodies adaptability, creativity, and a commitment to excellence both in my professional endeavors and personal passions.
More Speakers
Deep Learning (DL) adoption in industrial applications highlights the need for robust testing to ensure software quality and reliability. Unlike traditional software, DL lacks statistical guarantees and struggles with out-of-distribution (OOD) data. Current DL testing methods often overlook data distribution, leading to challenges in distinguishing genuine errors from OOD outliers. Addressing this gap, Case Study: Enhancing Deep Learning Software Quality Through OOD-Aware Testing Results from the study demonstrated significant improvements when utilizing OOD-aware testing techniques. Specifically, integrating data distribution awareness in testing and retraining processes led to a notable enhancement in error detection capabilities. Distribution-aware approaches outperformed traditional methods by up to 21.5%, highlighting their effectiveness in mitigating errors associated with OOD data and improving overall DL system performance. The case study underscores the critical importance of adopting OOD-aware testing strategies in DL software development. By enhancing the ability to identify and address errors caused by OOD data, organizations can significantly improve the reliability and performance of DL applications in real-world industrial settings.
