Audience: Faculty

This dissertation investigates how neuromorphic and brain-inspired principles can guide the design of efficient neural language architectures. It addresses two central limitations of modern Transformer-based language models: memory growth with context length and high computational cost from dense matrix multiplication. Through studies of spiking neural networks, linear-recurrent language models, hybrid attention architectures, MatMul-free models, and…

Sequencing technologies have revolutionized our understanding of biology, yet many existing methods require fragmentation of DNA or RNA, fundamentally limiting our ability to study these molecules in their native, intact forms. Long-read sequencing overcomes this constraint by enabling the sequencing of long, single-molecule native DNA and RNA, providing simultaneous access to both sequence and base…

As quantum computing transitions from theory to practice, understanding which algorithms suit near-term devices becomes critical. Current quantum computers are severely constrained by limited qubit counts, short coherence times, and high error rates that quickly degrade computation into noise. This thesis addresses two interconnected questions: what non-trivial computational tasks can near-term devices execute and how…

Centromeric DNA is critical for accurate chromosome segregation and genome stability, but due to its repetitive nature, it was only recently fully included in a human reference. Rapid evolution and sequence diversity in these regions limit the utility of one reference sequence, however. Integrating centromeric and pericentromeric satellite DNA – which together constitute over 5%…