AUTOMATED ASSESSMENT OF EXAMINATION SCRIPTS
First Claim
1. A computer-implemented method for classifying examination scripts, the method comprising:
- by a computing device, generating an aggregate vector for a set of input training samples, the aggregate vector normalized by a timing factor;
by the computing device, adjusting a weight vector based on the aggregate vector;
by the computing device, reducing the timing factor; and
by the computing device, repeating the generating, adjusting and reducing operations to generate an optimized weight vector.
2 Assignments
0 Petitions
Accused Products
Abstract
Embodiments herein provide automated assessment of examination scripts, such as English for Speakers of Other Languages (ESOL) examination scripts, written in response to prompts eliciting free text answers. In an embodiment, the task may be defined as discriminative preference ranking. Further, a system employing such methodology may be trained and tested on a corpus of manually-graded scripts. Embodiments herein, unlike extant solutions, are relatively prompt-insensitive and resistant to subversion, even if the operating principles are known. Embodiments may also detect scripts which are linguistically good but non-responsive to prompts such as memorized responses.
-
Citations
42 Claims
-
1. A computer-implemented method for classifying examination scripts, the method comprising:
-
by a computing device, generating an aggregate vector for a set of input training samples, the aggregate vector normalized by a timing factor; by the computing device, adjusting a weight vector based on the aggregate vector; by the computing device, reducing the timing factor; and by the computing device, repeating the generating, adjusting and reducing operations to generate an optimized weight vector. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 35)
-
-
27. A system for classifying examination scripts, the system comprising:
-
a processor; and a discriminative script classifier generator configured to operate the processor to; receive one or more training samples obtained from feature analysis of one or more training scripts; and generate a discriminative classifier from the training scripts. - View Dependent Claims (28, 29, 30, 31, 32, 33, 34)
-
-
36. One or more non-transitory computer-readable media containing instructions which, as a result of execution by a computing device, cause the computing device to perform a method, the method comprising:
-
generating an aggregate vector for a set of input training samples, the aggregate vector normalized by a timing actor; adjusting a weight vector based on the aggregate vector; reducing the timing factor; and repeating the generating, adjusting and reducing operations to generate an optimized weight vector. - View Dependent Claims (37, 38, 39, 40, 41, 42)
-
Specification