Sperm whales (Physeter macrocephalus) have followed fishing vessels off the Alaskan coast for decades, in order to remove sablefish ("depredate") from longlines. The Southeast Alaska Sperm Whale Avoidance Project (SEASWAP) has found that whales respond to distinctive acoustic cues made by hauling fishing vessels, as well as to marker buoys on the surface. Between 15-17 August 2010 a simple two-element vertical array was deployed off the continental slope of Southeast Alaska in 1200 m water depth. The array was attached to a longline fishing buoyline at 300 m depth, close to the sound-speed minimum of the deep-water profile. The buoyline also served as a depredation decoy, attracting seven sperm whales to the area. One animal was tagged with both a LIMPET dive depthtransmitting satellite and bioacoustic B-probe tag. Both tag datasets were used as an independent check of a passive acoustic scheme for tracking the whale in depth and range, which exploited the elevation angles and relative arrival times of multiple ray paths recorded on the array. The localization approach doesnt require knowledge of the local bottom bathymetry. Numerical propagation models yielded accurate locations up to at least 35 km range at Beaufort sea state 3. Ongoing work includes combining the arrival angle information with an algorithm developed by Le Bot et al. [1] that uses the rhythmic properties of odontocet click trains to separate interleaved click trains. This approach will improve our localization capabilities in presence of multiple sperm whales. In order to achieve better separation of interleaved click trains it is possible to use machine learning based algorithms. This new concept is based on finding useful information hidden in a large database. This useful information can then be represented by a sparse subspace. The first step of the approach is to extract informative features with a new detector proposed by Dadouchi et al. [2]. Once the dictionary of features is learned, any signal of this considered dataset can be approximated sparsely. By reducing the dimensional space, the sparse representation has the advantage to provide an optimally representation of the data. [Work supported by the North Pacific Research Board, the Alaska SeaLife Center, ONR, NOAA and ANR-12-ASTR-0021-03 "MER CALME"]