# US 2021/0241127 A1 · Amplifying group intelligence by adaptive population optimization

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2021/0241127 A1
- Document ID: US20210241127A1
- Status: Published application
- Kind: A1
- Application number: 17/237,972
- Filed: 2021-04-22
- Published: 2021-08-05
- Earliest claimed: 2014-03-26
- Assignees: Unanimous A. I., Inc.
- Applicants: Unanimous A. I., Inc.
- Inventors: LOUIS B. ROSENBERG; GREGG WILLCOX
- Web page: https://patentgrep.com/patent/US20210241127A1

## CPC

- G06N 5/022: Knowledge engineering; Knowledge acquisition
- G06F 16/95: Retrieval from the web
- G06F 17/18: for evaluating statistical data {, e.g. average values, frequency distributions, probability functions, regression analysis (forecasting specially adapted for a specific administrative, business or logistic context G06Q10/04)}
- G06N 3/006: based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
- G06N 5/041: {Abduction}
- G06N 20/00: Machine learning
- G06N 20/10: using kernel methods, e.g. support vector machines [SVM]
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

System and method for amplifying the accuracy of forecasts generated by software systems that harness the collective intelligence of human populations by curating optimized sub-populations through an intelligent selection process. Participants predict event outcomes and/or provide evaluations of their confidence in their predictions. The system determines a score wherein the alignment score indicates how well that participant's prediction aligns with the predictions given by the baseline population. Participants can then be selected from the population based on the participant alignment scores.

## Claims

### Claim 1

A system for curating an optimized population of human forecasting participants from a baseline population of human forecasting participants, the system comprising:
an analysis processing device including a processor and configured for network communication; and
a plurality of application instances, wherein each application instance is configured for interacting with one participant and is in network communication with the analysis processing device, wherein the system is configured to perform the steps of:
present, to each baseline population participant via the application instance, a set of possible outcomes for each of a set of future events, wherein each set of possible outcomes for each future event has at least two possible outcomes;
collect a set of predictions from each baseline population participant via the application instance, wherein each set of predictions includes, for each future event in the set of future events, one predicted outcome selected from the set of possible outcomes for that future event, and wherein an actual outcome of each of the future events is unknown at the time of collecting;
store the set of predictions from each baseline population participant;
for each future event in the set of future events, compute, using the stored sets of predictions, at least one outcome support value for each possible outcome of that future event, wherein each outcome support value is based at least in part on a percentage of baseline population participants that predicted that possible outcome for that future event;
for each baseline population participant, compute a set of alignment scores comprising an alignment score for each future event, wherein the alignment score for the future event is computed by algorithmically comparing the participant's predicted outcome for that future event to the outcome support value for that possible outcome of that future event, whereby the alignment score indicates how well that participant's prediction aligned with the predictions collected from the baseline population for that future event;
for each baseline population participant, determine an alignment index based on the set of alignment scores computed for that participant, wherein the alignment index indicates how well the set of that participant's predictions aligned with the predictions collected from the baseline population; and
curate an updated population of participants, the curating including at least one selected from the group of (a) culling a plurality of participants from the baseline population in response to the alignment index of each culled participant indicating low alignment compared to the baseline population, and (b) applying a weighting value to each of a set of baseline population participants, wherein the applied weighting values are lower for participants with the alignment index indicating low alignment with the baseline population and the applied weighting values are higher for baseline population participants with the alignment index indicating high alignment with the baseline population.

### Claim 2 (depends on 1)

The system for curating the optimized population of participants of claim 1, further comprising the system configured to perform the step of:
inviting users to participate, wherein the baseline population of participants comprises users who accept the invitation.

### Claim 3 (depends on 1)

The system for curating the optimized population of participants of claim 1, further comprising the system configured to perform the step of:
collecting personal information from the baseline population of participants.

### Claim 4 (depends on 1)

The system for curating the optimized population of participants of claim 1, further comprising the system configured to perform the step of:
after determining the alignment index for each baseline population participant, selecting a sub-population of participants based at least in part on the alignment indexes.

### Claim 5 (depends on 1)

The system for curating the optimized population of participants of claim 1, further comprising the system configured to perform the step of:
collecting, from each baseline population participant, at least one quantitative confidence assessment, wherein each confidence assessment represents that participant's confidence in their predicted outcome for one future event of the set of future events.

### Claim 6 (depends on 5)

The system for curating the optimized population of participants of claim 5, the system further configured to perform the step of:
weighting at least one outcome support value based on a plurality of the confidence assessments.

### Claim 7 (depends on 1)

The system for curating the optimized population of participants of claim 1, further comprising the system configured to perform the step of:
collecting, from each baseline population participant, a quantitative assessment of that participant's confidence in their knowledge of a specific knowledge category.

### Claim 8 (depends on 1)

The system for curating the optimized population of participants of claim 1, further comprising the system configured to perform the step of:
after curating the baseline population of participants, participating, by the curated population, in at least one crowd-based or swarm-based prediction session for a future event having at least two outcomes.

### Claim 9

A method for curating an optimized population of human forecasting participants from a baseline population of human forecasting participants, comprising the steps of:
presenting, by an analysis processing device including a processor and configured for networked communication, to each baseline population participant via an application instance in network communication with the analysis processing device and configured to interact with that participant, a set of possible outcomes for each of a set of future events, wherein each set of possible outcomes for each future event has at least two possible outcomes;
collecting a set of predictions from each baseline population participant via the application instance, wherein each set of predictions includes, for each future event in the set of future events, one predicted outcome selected from the set of possible outcomes for that future event, and wherein an actual outcome of each of the future events is unknown at the time of collecting;
storing the set of predictions from each baseline population participant;
computing, by the analysis processing device for each future event in the set of future events, at least one outcome support value for each possible outcome of that future event, wherein each outcome support value is based at least in part on a percentage of baseline population participants that predicted that possible outcome for that future event;
computing, by the analysis processing device for each baseline population participant, a set of alignment scores comprising an alignment score for each future event, wherein the alignment score for the future event is computed by algorithmically comparing the participant's predicted outcome for that future event to the outcome support value for that possible outcome of that future event, whereby the alignment score indicates how well participant's prediction aligned with the predictions collected from the baseline population for that future event;
determining, by the analysis processing device for each baseline population participant, an alignment index based on the set of alignment scores computed for that participant, wherein the alignment index indicates how well the set of that participant's predictions aligned with the predictions collected from the baseline population; and
curating, by the analysis processing device, an updated population of participants, based at least in part upon a plurality of the alignment indexes, the curating including at least one selected from the group of (a) culling a plurality of participants from the baseline population in response to the alignment index of each culled participant indicating low alignment compared to the baseline population, and (b) applying a weighting value to each of a set of baseline population participants, wherein the applied weighting values are lower for participants with the alignment index indicating low alignment with the baseline population and the applied weighting values are higher for baseline population participants with the alignment index indicating high alignment with the baseline population.

### Claim 10 (depends on 9)

The method for curating the optimized population of participants of claim 9, further comprising the step of:
inviting users to participate, wherein the baseline population of participants comprises users who accept the invitation.

### Claim 11 (depends on 9)

The method for curating the optimized population of participants of claim 9, further comprising the step of:
collecting, via the application instances, personal information from the baseline population of participants.

### Claim 12 (depends on 9)

The method for curating the optimized population of participants of claim 9, further comprising the step of:
selecting, by the analysis processing device after determining the alignment index for each baseline population participant, a sub-population of participants based at least in part on the alignment indexes.

### Claim 13 (depends on 9)

The method for curating the optimized population of participants of claim 9, further comprising the step of:
collecting, via the application instances from each participant, at least one quantitative confidence assessment, wherein each confidence assessment represents that participant's confidence in their predicted outcome for one future event of the set of future events.

### Claim 14 (depends on 13)

The method for curating the optimized population of participants of claim 13, further comprising the step of:
weighting at least one outcome support value based on a plurality of the confidence assessments.

### Claim 15 (depends on 9)

The method for curating the optimized population of participants of claim 9, further comprising the step of:
collecting, via the application instances for each baseline population participant, a quantitative assessment of that participant's confidence in their knowledge of a specific knowledge category.

### Claim 16 (depends on 9)

The method for curating the optimized population of participants of claim 9, further comprising the step of:
after curating the baseline population of participants, participating, by the curated population, in at least one crowd-based or swarm-based prediction session for a future event having at least two outcomes.