Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Hybrid Document Update #1617

Merged
merged 2 commits into from
Sep 1, 2023
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
133 changes: 124 additions & 9 deletions doc/Hybrid.lyx
Original file line number Diff line number Diff line change
Expand Up @@ -86,7 +86,7 @@ Hybrid Inference
\end_layout

\begin_layout Author
Frank Dellaert & Varun Agrawal
Frank Dellaert
\end_layout

\begin_layout Date
Expand Down Expand Up @@ -264,12 +264,110 @@ If we take the log of
we get
\begin_inset Formula
\begin{equation}
\log P(x|y,m)=\log P_{m}(x|y)=K_{gcm}-E_{gcm}(x,y).\label{eq:gm_log}
\log P(x|y,m)=\log P_{m}(x|y)=K_{gc}(m)-E_{gcm}(x,y).\label{eq:gm_log}
\end{equation}

\end_inset

Equating

\end_layout

\begin_layout Standard
\noindent
For conciseness, we will write
\begin_inset Formula $K_{gc}(m)$
\end_inset

as
\begin_inset Formula $K_{gcm}$
\end_inset

.
\end_layout

\begin_layout Standard
\SpecialChar allowbreak

\end_layout

\begin_layout Standard
\noindent
The key point here is that
\family roman
\series medium
\shape up
\size normal
\emph off
\bar no
\strikeout off
\xout off
\uuline off
\uwave off
\noun off
\color none

\begin_inset Formula $K_{gm}$
\end_inset


\family default
\series default
\shape default
\size default
\emph default
\bar default
\strikeout default
\xout default
\uuline default
\uwave default
\noun default
\color inherit
is the log-normalization constant for the complete
\emph on
GaussianMixture
\emph default
across all values of
\begin_inset Formula $m$
\end_inset

, and cannot be dependent on the value of
\begin_inset Formula $m$
\end_inset

.
In contrast,
\begin_inset Formula $K_{gcm}$
\end_inset

is the log-normalization constant for a specific
\emph on
GaussianConditional
\emph default
mode (thus dependent on
\begin_inset Formula $m$
\end_inset

) and can have differing values based on the covariance matrices for each
mode.
Thus to obtain a constant
\begin_inset Formula $K_{gm}$
\end_inset

which satisfies the invariant, we need to specify
\begin_inset Formula $E_{gm}(x,y,m)$
\end_inset

accordingly.
\end_layout

\begin_layout Standard
\SpecialChar allowbreak

\end_layout

\begin_layout Standard
\noindent
By equating
\begin_inset CommandInset ref
LatexCommand eqref
reference "eq:gm_invariant"
Expand All @@ -289,7 +387,7 @@ noprefix "false"

\end_inset

we see that this can be achieved by defining the error
, we see that this can be achieved by defining the error
\begin_inset Formula $E_{gm}(x,y,m)$
\end_inset

Expand Down Expand Up @@ -541,9 +639,9 @@ noprefix "false"

:
\begin_inset Formula
\[
E_{mf}(y,m)=C+E_{gcm}(\bar{x},y)-K_{gcm}.
\]
\begin{equation}
E_{mf}(y,m)=C+E_{gcm}(\bar{x},y)-K_{gcm}\label{eq:mixture_factor}
\end{equation}

\end_inset

Expand All @@ -555,8 +653,25 @@ Unfortunately, we can no longer choose
\begin_inset Formula $m$
\end_inset

to make the constant disappear.
There are two possibilities:
to make the constant disappear, since
\begin_inset Formula $C$
\end_inset

has to be a constant applicable across all
\begin_inset Formula $m$
\end_inset

.
\end_layout

\begin_layout Standard
\SpecialChar allowbreak

\end_layout

\begin_layout Standard
\noindent
There are two possibilities:
\end_layout

\begin_layout Enumerate
Expand Down
Binary file modified doc/Hybrid.pdf
Binary file not shown.