APPGATE Brainspace User Manual

Download

Page 1

Brainspace

Preserve Expert Decision IP with Portable Learning

Reduce the machine learning

curve with Brainspace’s

Portable Learning features.

Bend the Machine

Learning Curve

Brainspace’s Continuous Multimodal Learning (CMML) has revolutionized the application of supervised machine learning in litigation, investigations, and beyond. CMML tightly integrates machine learning with document tagging and Notebooks to capture relevant documents

as they are identied. The positive and

negative examples can then be exploited

for training with no extra effort. Classiers

can be trained for as many topics as are of interest, either natively in Brainspace

or in a review platform. Classiers can

then be combined with Brainspace’s other analytics in a synergistic cycle.

Preserve Your

Decision IP

Encode Expert

Knowledge

Portable Learning takes CMML to the next level. Now you can reuse and update your trained models, moving them from data set to data set and growing their effectiveness and

breadth each time. Portable classiers

can be inspected and edited as well, or even created manually from scratch. Preserve Decision IP by building

predictive models. Then effortlessly

transfer those models between datasets, generating nearly instant

relevance identication results on

new cases without the overhead of

searching and ltering.

Repeatable

Flashes of Insight

DON’T START FROM ZERO

STAT E

Adapting existing

classiers to a new data

set greatly reduces the

coding effort, and time, to

get classication results.

Brainspace’s leading edge

active learning reduces that

effort even more, allowing

rapid tuning of existing

classiers to new data sets.

APPGATE.COM

REUSE WHAT YOU’VE

LEARNED

Created a predictive model

that identies specic

behavior or sentiment?

Re-use it with Portable

Learning. Carry it to new

data sets, and build its power over time. Add

value to every project by

developing a library of

effective reusable classiers

for domains of interest.

LEVERAGE YOUR EXPERTS

Incorporate your team’s domain knowledge in a

classier, then update it (or

not) with machine learning.

Even create a predictive

model from scratch from a

list of keywords.

CAREFULLY TUNED

REUSABLE MODE LS

Portable models provide

a set of carefully tuned alterative perspectives

through which to view each

new data set. By building

processes around portable

models, investigatory

techniques can be applied

systematically at scale to

ensure nothing falls through

the crack.