Optimal Synthesis of Robust IDK Classifier Cascades

TY - JOUR

T1 - Optimal Synthesis of Robust IDK Classifier Cascades

AU - Baruah, Sanjoy

AU - Burns, Alan

AU - Davis, Robert Ian

N1 - © 2023 Copyright held by the owner/author(s).

PY - 2023/9/9

Y1 - 2023/9/9

N2 - An IDK classifier is a computing component that categorizes inputs into one of a number of classes, if it is able to do so with the required level of confidence, otherwise it returns “I Don’t Know” (IDK). IDK classifier cascades have been proposed as a way of balancing the needs for fast response and high accuracy in classification-based machine perception. Efficient algorithms for the synthesis of IDK classifier cascades have been derived; however, the responsiveness of these cascades is highly dependent on the accuracy of predictions regarding the run-time behavior of the classifiers from which they are built. Accurate predictions of such run-time behavior is difficult to obtain for many of the classifiers used for perception. By applying the algorithms using predictions framework, we propose efficient algorithms for the synthesis of IDK classifier cascades that are robust to inaccurate predictions in the following sense: the IDK classifier cascades synthesized by our algorithms have short expected execution durations when the predictions are accurate, and these expected durations increase only within specified bounds when the predictions are inaccurate.

AB - An IDK classifier is a computing component that categorizes inputs into one of a number of classes, if it is able to do so with the required level of confidence, otherwise it returns “I Don’t Know” (IDK). IDK classifier cascades have been proposed as a way of balancing the needs for fast response and high accuracy in classification-based machine perception. Efficient algorithms for the synthesis of IDK classifier cascades have been derived; however, the responsiveness of these cascades is highly dependent on the accuracy of predictions regarding the run-time behavior of the classifiers from which they are built. Accurate predictions of such run-time behavior is difficult to obtain for many of the classifiers used for perception. By applying the algorithms using predictions framework, we propose efficient algorithms for the synthesis of IDK classifier cascades that are robust to inaccurate predictions in the following sense: the IDK classifier cascades synthesized by our algorithms have short expected execution durations when the predictions are accurate, and these expected durations increase only within specified bounds when the predictions are inaccurate.

KW - real-time

KW - classifier

KW - IDK cascades

KW - synthesis

UR - https://www-users.york.ac.uk/~rd17/

U2 - https://doi.org/10.1145/3609129

DO - https://doi.org/10.1145/3609129

M3 - Article

SN - 1539-9087

VL - 22

SP - 1

EP - 26

JO - ACM Transactions in Embedded Computing Systems

JF - ACM Transactions in Embedded Computing Systems

IS - 5

M1 - 150

ER -