[TD‐P‐015]: LUDWIG: A CONVERSATIONAL ROBOT FOR PEOPLE WITH ALZHEIMER'S
Bibliographic record
Abstract
Assistive robotics are being produced to improve the quality-of-life for people with Alzheimer's disease (AD), including aid in activities of daily living. Recently, computer programs based on the statistics of human dialogue have been shown to reduce the cost of hand-crafting complex dialogue systems. Of these, one increasingly popular approach is the partially-observable Markov decision process (POMDP). Here, dialogue is represented as statistical relationships between observations (i.e., the words spoken and other linguistic/acoustic aspects of speech), unknown states (e.g., the participant's current level of confusion), and actions that can be performed (i.e., the computer's own speech output). We have recently built POMDPs to detect and avoid trouble-indicating behaviors (TIBs), with 82% and 96.1% accuracy, respectively, in the speech of people with AD. We are developing software that uniquely embodies TIB detection, avoidance, and recovery in dialogue within an on-line assistive robot, called Ludwig. Twelve individuals with AD have been recruited from a clinical care partner. In each of three sessions, participants describe a series of photographs shown on a tablet computer, and answer questions posed by an interviewer. In the first session, the interviewer is a human volunteer. The second session replaces them with the robot Ludwig, controlled remotely. In the third session, we use a completely automated system based on recent work to avoid confusion. At the time of the conference, all three sessions will have been completed for each participant. Preliminary qualitative analysis indicates broad interest in the project, and in Ludwig specifically, but issues related to attention and executive function in participants remain a challenge. Success with the task varies directly with the cognitive status of the older adult with AD, as measured by a mini-mental state exam. Future work must focus on extra-linguistic means of interaction to focus the attention of the older adult with AD. 1. Brodaty H, Connors MH, Xu J, Woodward M, Ames D, Group PS. The course of neuropsychiatric symptoms in dementia: A 3-year longitudinal study. Journal of the American Medical Directors Association. 2015;16(5):380–387. 2. Sadak TI, Katon J, Beck C, Cochrane BB, Borson S. Key neuropsychiatric symptoms in common dementias: prevalence and implications for caregivers, clinicians, and health systems. Research in gerontological nursing. 2014;7(1):44. 3. Steinberg M, Shao H, Zandi P, et al. Point and 5-year period prevalence of neuropsychiatric symptoms in dementia: the Cache County Study. International journal of geriatric psychiatry. 2008;23(2):170. 4. Gitlin LN, Kales HC, Lyketsos CG. Nonpharmacologic management of behavioral symptoms in dementia. JAMA. 2012;308(19):2020–2029. 5. Kales HC, Gitlin LN, Lyketsos CG. State of the Art Review: Assessment and management of behavioral and psychological symptoms of dementia. BMJ: British Medical Journal. 2015;350.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.057 | 0.020 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".