Article Instance

HTTP 200 OK
Allow: GET
Content-Type: application/json
Vary: Accept

{
    "pk": 27768,
    "title": "Can a Recurrent Neural Network Learn to Count Things?",
    "subtitle": null,
    "abstract": "We explore a recurrent neural network model of counting\nbased on the differentiable recurrent attentional model of\nGregor et al. (2015). Our results reveal that the model can\nlearn to count the number of items in a display, pointing to each\nof the items in turn and producing the next item in the count\nsequence at each step, then saying ‘done’ when there are no\nmore blobs to count. The model thus demonstrates that the\nability to learn to count does not depend on special knowledge\nrelevant to the counting task. We find that the model’s ability\nto count depends on how well it has learned to point to each\nsuccessive item in the array, underscoring the importance of\ncoordination of the visuospatial act of pointing with the\nrecitation of the count list. The model learns to count items in\na display more quickly if it has previously learned to touch all\nthe items in such a display correctly, capturing the relationship\nbetween touching and counting noted by Alibali and DiRusso.\nIn such cases it achieves performance sometimes thought to\nresult from a semantic induction of the ‘cardinality principle’.\nYet the errors that it makes have similarities with the patterns\nseen in human children’s counting errors, consistent with idea\nthat children rely on graded and somewhat variable\nmechanisms similar to our neural networks.",
    "language": "eng",
    "license": {
        "name": "",
        "short_name": "",
        "text": null,
        "url": ""
    },
    "keywords": [
        {
            "word": "Mathematical cognition"
        },
        {
            "word": "numerical cognition"
        },
        {
            "word": "Neural Networks"
        },
        {
            "word": "development"
        },
        {
            "word": "learning"
        },
        {
            "word": "transfer learning"
        }
    ],
    "section": "Publication-based-Talks",
    "is_remote": true,
    "remote_url": "https://escholarship.org/uc/item/33h3496v",
    "frozenauthors": [
        {
            "first_name": "Mengting",
            "middle_name": "",
            "last_name": "Fang",
            "name_suffix": "",
            "institution": "Beijing Normal University",
            "department": ""
        },
        {
            "first_name": "Zhenglong",
            "middle_name": "",
            "last_name": "Zhou",
            "name_suffix": "",
            "institution": "John Hopkins",
            "department": ""
        },
        {
            "first_name": "Sharon",
            "middle_name": "Y",
            "last_name": "Chen",
            "name_suffix": "",
            "institution": "Columbia University",
            "department": ""
        },
        {
            "first_name": "James",
            "middle_name": "L",
            "last_name": "McClelland",
            "name_suffix": "",
            "institution": "Stanford",
            "department": ""
        }
    ],
    "date_submitted": null,
    "date_accepted": null,
    "date_published": "2018-01-01T18:00:00Z",
    "render_galley": null,
    "galleys": [
        {
            "label": "PDF",
            "type": "pdf",
            "path": "https://journalpub.escholarship.org/cognitivesciencesociety/article/27768/galley/17408/download/"
        }
    ]
}