1 hr ago
Experts Urge Clean, Diverse Data for More Reliable AI
AI systems learn from data, much like students learn from examples.
Professor Nilanjan Dey said that having more data is not always enough.
The data must also be accurate, clean, and varied.
If an AI system sees only repeated examples, it may struggle with new situations.
Data from many different real-world conditions can help AI work better.
Dey said AI could help healthcare workers manage their workload.
He noted reports estimating that doctors in India spend an average of 2.8 minutes with each patient, compared with 48 seconds in Bangladesh.
He also said synthetic data may influence the next stage of AI development.
Professor Nilanjan Dey warned that unreliable or unscientifically collected data can produce flawed AI results.
He said increasing data volume alone does not necessarily improve artificial intelligence systems.
Diverse data covering more variables and real-world situations can make models more robust.
Limited or repetitive training datasets may cause models to perform poorly in unfamiliar situations.
Dey highlighted healthcare applications, synthetic data, and technology that could reduce pressure on doctors.
- Who
- Professor Nilanjan Dey of Techno International New Town.
- What
- He argued that reliable, clean, and diverse data is more important than simply collecting larger amounts of data for AI systems.
- Where
- AU Engineering College in Visakhapatnam.
- When
- Wednesday, during the ongoing international conference QUIST-A.
- Why
- To explain how data quality and diversity affect AI reliability, including potential healthcare applications.
Key facts
- Speaker
- Professor Nilanjan Dey
- Conference
- International conference Quantum Enhanced Sustainable Technologies for AI Systems (QUIST-A)
- Venue
- AU Engineering College, Visakhapatnam
- Core message
- More data does not necessarily mean better AI.
- Recommended focus
- Clean data and variable data rather than only big data
- Healthcare statistic
- A recent report cited average doctor-patient times of 2.8 minutes in India and 48 seconds in Bangladesh
- Emerging development
- Synthetic data could shape the next phase of AI
Quotes
Professor Nilanjan Dey
Professor at Techno International New Town who spoke at the QUIST-A conference
“More data does not necessarily mean better AI”
thehansindia.com
Sources
Related news
AI Could Bridge Women's Health Gap, But Better Data Key
AI won't scale without trusted data, says Infosys executive
AI May Bring Specialist Care to Rural India, Training Lags
India Bets Human Motion Data Can Teach Robots
Clean Centralised Data Seen as Key to E-Commerce AI Success
Cloudera CEO Says India Is Growth Market and R&D Hub
Bangladesh Recall Shanto, Mahmud for Asian Games T20 Squad
Sharma and Malik Share Halfway Lead at Golconda Masters
India Replaces Sri Lanka as 2027 Women’s Champions Trophy Host
Centre Assures JDU Bihar’s Interests In Farakka Treaty Decision
