Projects

Machine Learning

• Scene Classification — into different categories.
• Cutensor — Tensor decompositions in Torch. Application : TensorFaces
• Metric Learning — Implementation of different metric learning algorithms.
• Nyaya - Low level GPU API for Machine learning workloads.
• Classification of singing voice and instrumental sounds - Using SVM for classification.
• Cloud Tracking - scale space classification for image segmentation. Cloud Motion Vector is generated for prediction.

General

• Programming Primitives - Data structures + algorithms => programs
• Lossless Compression- Lossless compression pipeline with various algorithms and methodology.
  • LZW data compression
  • Image compression

Revisiting Earlier Projects

• CPU + embeddGPU - Platform for architectural design space exploration.
• Stream buffers - Trace driven simulation of stream buffers for data cache prefetching.
• Range Compression - implementation of range compression algorithm on GPUs.
• Tiger Compiler - compiler developed in Standard ML for toy programming laguage Tiger.
• Sorting Algorithms - Evaluation of different sorting algorithms.
• Lossless Image compression - Wavelet transform in Integer domain for Compression.