=== Emma Robinson - searches for stochastic backgrounds ===

- astrophysical and cosmological stochastic backgrounds
- can form noise-orthogonal TDI observables; but the all-sky overlap reduction function is zero

- construct TDI variable T (GW response highly suppressed)
- use T to discriminate the GWs in A and E
- assume exact suppression in T

- Bayesian analysis applied to MLDC 2.1 data
- variance variables

- to do in 3.5

=== N. Kanda - Numerical trial of cleaning of gravitational wave foreground from neutron star binaries in DECIGO ===

- 10^4-6 binary mergers in DECIGO: confusion noise around 0.1 Hz
	- binary detection will reach a few Gpc
- huge number of templates needed for signal identification

- simulated DECIGO noise and NS-NS foreground
- "tagging": using tails of binaries
- subtraction: a priori identification + errors on parameters

- 700 Mpc cleaning removes 30% of spectrum (900 Mpc -> 60%)

=== Antoine Petiteau - Searching for spinning supermassive black hole binaries using a genetic algorithm ===

- Genetic algorithm: initial state -> selection -> breeding -> mutation -> repeat
- Genes = parameters; organism = waveform; quality = maximized likelihood
- Roulette selection (probability proportional to quality) of two breeders
	- Mix genes of parents
	- Change few bits with probability mutation rate
- Acceleration processes
	- Elitism (cloning)
	- Simulated annealing
	- Evolution of PMR

- Works well for nonspinning SMBH binaries and multiple sources

=== Ed Porter - Searching for spinning SMBHs in the 3rd Mock LISA Data Challenge ===

- parallel tempering
- have successfully detected all sources in the traning data sets
- need to improve parameter estimation