Educational Testing Service

Derrick Higgins

Klaus Zechner

Michael Heilman

Lei Chen

Jill C Burstein

Nitin Madnani

Yoko Futagi

Daniel Marcu

Keelan Evanini

Su-Youn Yoon

Isaac I Bejar

Shasha Xie

Rene R Lawless

Je Hun Jeon

Claudia Gentile

Michael Flor

Paul Douglas Deane

Guangming Ling

Systems and methods are provided for automatically scoring a constructed response. The constructed response is processed to generate a plurality of numerical vectors that is representative of the constructed response. A model is applied to the plurality of numerical vectors. The model includes an input layer configured to receive the plurality of numerical vectors, the input layer being connected to a following layer of the model via a first plurality of connections. Each of the connections has a first weight. An intermediate layer of nodes is configured to receive inputs from an immediately-preceding layer of the model via a second plurality of connections, each of the connections having a second weight. An output layer is connected to the intermediate layer via a third plurality of connections, each of the connections having a third weight. The output layer is configured to generate a score for the constructed response.

Deep convolutional neural networks for automated scoring of constructed responses

A method and system for using a natural language generator for automatic assessment item generation is disclosed. The natural language generator includes a document structure generator that produces an abstract document specification defining a structure for an assessment item based on user input. The abstract document specification is input into a logical schema generator, which produces a logical schema specification that creates a more detailed specification for an assessment item. Finally, a sentence generator receives the logical schema specification and creates natural language for the assessment item based on the variables defined in the logical schema specification.

Automated test item generation system and method

Systems and methods are provided for determining a susceptibility of a computer-implemented automated scoring engine to gaming strategies. A plurality of responses to a prompt are provided to a computer-implemented automated scoring engine to receive a first set of scores. A first transformation is performed on each of the plurality of responses to generate a first set of transformed responses. The first set of transformed responses is provided to the computer-implemented automatic scoring engine to receive a second set of scores, and a gaming susceptibility metric is determined based on the first set of scores and the second set of scores.

Systems and methods for evaluating susceptibility of automated scoring engines to gaming behavior

A method for scoring non-native speech includes receiving a speech sample spoken by a non-native speaker and performing automatic speech recognition and metric extraction on the speech sample to generate a transcript of the speech sample and a speech metric associated with the speech sample. The method further includes determining whether the speech sample is scorable or non-scorable based upon the transcript and speech metric, where the determination is based on an audio quality of the speech sample, an amount of speech of the speech sample, a degree to which the speech sample is off-topic, whether the speech sample includes speech from an incorrect language, or whether the speech sample includes plagiarized material. When the sample is determined to be non-scorable, an indication of non-scorability is associated with the speech sample. When the sample is determined to be scorable, the sample is provided to a scoring model for scoring.

Non-scorable response filters for speech scoring systems

A computer-implemented method, system, and computer program product for generating vector-based similarity scores in text document comparisons considering confounding effects of document length. Vector-based methods for comparing the semantic similarity between texts (such as Content Vector Analysis and Random Indexing) have a characteristic which may reduce their usefulness for some applications: the similarity estimates they produce are strongly correlated with the lengths of the texts compared. The statistical basis for this confound is described, and suggests the application of a pivoted normalization method from information retrieval to correct for the effect of document length. In two text categorization experiments, Random Indexing similarity scores using pivoted normalization are shown to perform significantly better than standard vector-based similarity estimation methods.

System and method for handling the confounding effect of document length on vector-based similarity scores

The present disclosure presents a useful metric for assessing the relative difficulty which non-native speakers face in pronouncing a given utterance and a method and systems for using such a metric in the evaluation and assessment of the utterances of non-native speakers. In an embodiment, the metric may be based on both known sources of difficulty for language learners and a corpus-based measure of cross-language sound differences. The method may be applied to speakers who primarily speak a first language speaking utterances in any non-native second language.

Method and system for assessing pronunciation difficulties of non-native speakers

A method and system for determining text coherence in an essay is disclosed. A method of evaluating the coherence of an essay includes receiving an essay having one or more discourse elements and text segments. The one or more discourse elements are annotated either manually or automatically. A text segment vector is generated for each text segment in a discourse element using sparse random indexing vectors. The method or system then identifies one or more essay dimensions and measures the semantic similarity of each text segment based on the essay dimensions. Finally, a coherence level is assigned to the essay based on the measured semantic similarities.

Method and system for determining text coherence

A tool, method, and system for use in the development of sentence-based test items are disclosed. The tool may include a user interface that may include a database selection field, a sentence pattern entry field, an option pane, and an output pane. The tool may search a database for one or more sentences and may generate one or more responses to the one or more sentences. The one or more sentences and one or more responses may be used to produce the sentence-based test items. The tool may allow test items to be developed more quickly and easily than manual test item authoring. Accordingly, test item development costs may be lowered and test security may be enhanced.

Method and system for text retrieval for computer-assisted item creation

Method and system for assessing pronunciation difficulties of non-native speakers by entropy calculation

Methods and systems for detecting off-topic essays are described that do not require training using human-scored essays. The methods can detect different types of off-topic essays, such as unexpected topic essays and bad faith essays. Unexpected topic essays are essays that address an incorrect topic. Bad faith essays address no topic. The methods can use content vector analysis to determine the similarity between the essay and one or more prompts. If the essay prompt with which an essay is associated is among the most similar to the essay, the essay is on-topic. Otherwise, the essay is considered to be an unexpected topic essay. Similarly, if the essay is sufficiently dissimilar to all essay prompts, the essay is considered to be a bad faith essay.

Growing community of inventors

Highland Park, NJ, United States of America

Derrick Higgins