Weizmann Logo
ECCC
Electronic Colloquium on Computational Complexity

Under the auspices of the Computational Complexity Foundation (CCF)

Login | Register | Classic Style



REPORTS > DETAIL:

Revision(s):

Revision #3 to TR21-133 | 24th May 2023 10:15

Testing Distributions of Huge Objects

RSS-Feed




Revision #3
Authors: Oded Goldreich, Dana Ron
Accepted on: 24th May 2023 10:15
Downloads: 150
Keywords: 


Abstract:

We initiate a study of a new model of property testing that is a hybrid of testing properties of distributions and testing properties of strings.
Specifically, the new model refers to testing properties of distributions, but these are distributions over huge objects (i.e., very long strings).
Accordingly, the model accounts for the total number of local probes into these objects (resp., queries to the strings)as well as for the distance between objects (resp., strings).
Specifically, the distance between distributions is defined as the earth mover's distance with respect to the relative Hamming distance between strings.

We study the query complexity of testing in this new model, focusing on three directions.
First, we try to relate the query complexity of testing properties in the new model to the sample complexity of testing these properties in the standard distribution testing model.
Second, we consider the complexity of testing properties that arise naturally in the new model (e.g., distributions that capture random variations of fixed strings).
Third, we consider the complexity of testing properties that were extensively studied in the standard distribution testing model: Two such cases are uniform distributions
and pairs of identical distributions, where we obtain the following results.

\begin{itemize}
\item
Testing whether a distribution over $n$-bit long strings is uniform on some set of size $m$ can be tested with query complexity ${\widetilde O}(m/\epsilon^3)$, where $\epsilon>(\log_2m)/n$ is the proximity parameter.
\item
Testing whether two distribution over $n$-bit long strings that have support size at most $m$ are identical can be tested with query complexity ${\widetilde O}(m^{2/3}/\epsilon^3)$.
\end{itemize}
Both upper bounds are pretty tight; that is, for $\epsilon=\Omega(1)$, the first task requires $\Omega(m^c)$ queries for any $c<1$ and $n=\omega(\log m)$, whereas the second task requires $\Omega(m^{2/3})$ queries.
Note that the query complexity of the first task is higher than the sample complexity of the corresponding task in the standard distribution testing model, whereas in the case of the second task the bounds almost match.



Changes to previous version:

Improved/extended introduction


Revision #2 to TR21-133 | 23rd January 2022 18:12

Testing Distributions of Huge Objects





Revision #2
Authors: Oded Goldreich, Dana Ron
Accepted on: 23rd January 2022 18:12
Downloads: 244
Keywords: 


Abstract:

We initiate a study of a new model of property testing that is a hybrid of testing properties of distributions and testing properties of strings.
Specifically, the new model refers to testing properties of distributions, but these are distributions over huge objects (i.e., very long strings).
Accordingly, the model accounts for the total number of local probes into these objects (resp., queries to the strings)as well as for the distance between objects (resp., strings).
Specifically, the distance between distributions is defined as the earth mover's distance with respect to the relative Hamming distance between strings.

We study the query complexity of testing in this new model, focusing on three directions.
First, we try to relate the query complexity of testing properties in the new model to the sample complexity of testing these properties in the standard distribution testing model.
Second, we consider the complexity of testing properties that arise naturally in the new model (e.g., distributions that capture random variations of fixed strings).
Third, we consider the complexity of testing properties that were extensively studied in the standard distribution testing model: Two such cases are uniform distributions
and pairs of identical distributions, where we obtain the following results.

\begin{itemize}
\item
Testing whether a distribution over $n$-bit long strings is uniform on some set of size $m$ can be tested with query complexity ${\widetilde O}(m/\epsilon^3)$, where $\epsilon>(\log_2m)/n$ is the proximity parameter.
\item
Testing whether two distribution over $n$-bit long strings that have support size at most $m$ are identical can be tested with query complexity ${\widetilde O}(m^{2/3}/\epsilon^3)$.
\end{itemize}
Both upper bounds are pretty tight; that is, for $\epsilon=\Omega(1)$, the first task requires $\Omega(m^c)$ queries for any $c<1$ and $n=\omega(\log m)$, whereas the second task requires $\Omega(m^{2/3})$ queries.
Note that the query complexity of the first task is higher than the sample complexity of the corresponding task in the standard distribution testing model, whereas in the case of the second task the bounds almost match.



Changes to previous version:

Correcting inaccuracies in the stmt of Thm 1.6 (and 2.2), and correcting corresponding inaccuracies in the proof of Thm 2.2 and in Sec 2.3.


Revision #1 to TR21-133 | 16th September 2021 18:15

Testing Distributions of Huge Objects





Revision #1
Authors: Oded Goldreich, Dana Ron
Accepted on: 16th September 2021 18:15
Downloads: 375
Keywords: 


Abstract:

We initiate a study of a new model of property testing that is a hybrid of testing properties of distributions and testing properties of strings.
Specifically, the new model refers to testing properties of distributions, but these are distributions over huge objects (i.e., very long strings).
Accordingly, the model accounts for the total number of local probes into these objects (resp., queries to the strings)as well as for the distance between objects (resp., strings).
Specifically, the distance between distributions is defined as the earth mover's distance with respect to the relative Hamming distance between strings.

We study the query complexity of testing in this new model, focusing on three directions.
First, we try to relate the query complexity of testing properties in the new model to the sample complexity of testing these properties in the standard distribution testing model.
Second, we consider the complexity of testing properties that arise naturally in the new model (e.g., distributions that capture random variations of fixed strings).
Third, we consider the complexity of testing properties that were extensively studied in the standard distribution testing model: Two such cases are uniform distributions
and pairs of identical distributions, where we obtain the following results.

\begin{itemize}
\item
Testing whether a distribution over $n$-bit long strings is uniform on some set of size $m$ can be tested with query complexity ${\widetilde O}(m/\epsilon^3)$, where $\epsilon>(\log_2m)/n$ is the proximity parameter.
\item
Testing whether two distribution over $n$-bit long strings that have support size at most $m$ are identical can be tested with query complexity ${\widetilde O}(m^{2/3}/\epsilon^3)$.
\end{itemize}
Both upper bounds are pretty tight; that is, for $\epsilon=\Omega(1)$, the first task requires $\Omega(m^c)$ queries for any $c<1$ and $n=\omega(\log m)$, whereas the second task requires $\Omega(m^{2/3})$ queries.
Note that the query complexity of the first task is higher than the sample complexity of the corresponding task in the standard distribution testing model, whereas in the case of the second task the bounds almost match.



Changes to previous version:

Adding a new result (which appears as Thm 1.6)
and detailing the proof of Thm 5.2.


Paper:

TR21-133 | 12th September 2021 14:23

Testing Distributions of Huge Objects





TR21-133
Authors: Oded Goldreich, Dana Ron
Publication: 12th September 2021 14:24
Downloads: 502
Keywords: 


Abstract:

We initiate a study of a new model of property testing that is a hybrid of testing properties of distributions and testing properties of strings.
Specifically, the new model refers to testing properties of distributions, but these are distributions over huge objects (i.e., very long strings).
Accordingly, the model accounts for the total number of local probes into these objects (resp., queries to the strings)as well as for the distance between objects (resp., strings).
Specifically, the distance between distributions is defined as the earth mover's distance with respect to the relative Hamming distance between strings.

We study the query complexity of testing in this new model, focusing on three directions.
First, we try to relate the query complexity of testing properties in the new model to the sample complexity of testing these properties in the standard distribution testing model.
Second, we consider the complexity of testing properties that arise naturally in the new model (e.g., distributions that capture random variations of fixed strings).
Third, we consider the complexity of testing properties that were extensively studied in the standard distribution testing model: Two such cases are uniform distributions
and pairs of identical distributions, where we obtain the following results.

\begin{itemize}
\item
Testing whether a distribution over $n$-bit long strings is uniform on some set of size $m$ can be tested with query complexity ${\widetilde O}(m/\epsilon^3)$, where $\epsilon>(\log_2m)/n$ is the proximity parameter.
\item
Testing whether two distribution over $n$-bit long strings that have support size at most $m$ are identical can be tested with query complexity ${\widetilde O}(m^{2/3}/\epsilon^3)$.
\end{itemize}
Both upper bounds are pretty tight; that is, for $\epsilon=\Omega(1)$, the first task requires $\Omega(m^c)$ queries for any $c<1$ and $n=\omega(\log m)$, whereas the second task requires $\Omega(m^{2/3})$ queries.
Note that the query complexity of the first task is higher than the sample complexity of the corresponding task in the standard distribution testing model, whereas in the case of the second task the bounds almost match.



ISSN 1433-8092 | Imprint