A library of algorithms
You know, beautifully crafted algorithms play a central role in autonomic computing implementation -- resource monitoring, problem determination, deterministic action decisions, knowledge base accrual. IBM Research supports groups that develop mathematical ways of enabling these features. The researchers at the IBM Almaden Lab provide a list of papers on the algorithms they've collaborated on.
You'll find algorithms on
- Data streams (counting distinct elements, approximating higher frequency moments, allocating bnadwidth, computing massive datasets).
- Embeddings and metric spaces (navigating nets, measuring descent, doubling metrics).
- Approximations (minimum latency).
- Peer-to-peer networks (load balancing, object location).
- Aggregation (rank Web aggregation, instance optimal, similarity search, partial ranking).
- Sampling and sublinear time (testing independence, approximating entrophy).
- Lattice (shortest or closest vector determination).
Almaden also supports research into complexity theory in the following areas:
- Communication complexity (information theory, randomization vs. nondeterminism, quantum vs. classical one-way communication).
- Lower bounds in concrete models.
- Lattice problems (random and worst-case behaviors).
- Cryptography and coding.