The IceCube Neutrino Observatory has detected a diffuse astrophysical neutrino
flux which is expected to have a close to equal ratio of neutrino flavors due to
thorough mixing over astronomical baselines. However, IceCube has yet to detect
astrophysical tau neutrinos. A tau neutrino undergoing charged current interaction will
have two subsequent energy losses, the first from the neutrino interaction and the
second from the decay of the secondary tau lepton. At PeV neutrino energies, IceCube
can resolve these depositions as two separated cascades. At energies near hundreds of
TeV the cascades are not well separated but can be observed as a double pulse waveform
in individual IceCube sensors. Here we will present three techniques to improve tau double
pulse waveform identification. One technique utilizes neighboring IceCube light sensors to
observe a double pulse event, another incorporates machine learning algorithms to boost
detection of double pulses, and the third explores the possibility of stacking waveforms
to increase the double pulse signal. The first two approaches show a promising increase of
signal rates by at least 50\% while keeping similar or lower backgrounds at the double
pulse waveform identification stage.