Predicting Startup Funding Momentum with Collective Intelligence
Bibliographic record
Abstract
In this paper, we describe a set of techniques that leverage human and machine intelligence to create data and models that are both predictively accurate and support an in-depth explanation facility. Our approach leverages the cognitive diversity of an evaluation team with domain experience and investment experience. Each team, unique to each startup, participates in an interactive evaluation process where each team member uses available evidence to score the potential for a startup to produce a return on investment. All startups are then tracked for performance following scoring and prediction. To date, the approach outlined here has demonstrated high predictive accuracy.Key to the explanatory power of our approach is collecting natural language data on reasons behind a particular score. All supporting reasons for all scores are collected and sampled to learn the rank and relevance determined by group responses. Each reason is assigned a relevancy score based on structured interactions with the evaluating team. A typical evaluation will consist of ~200 quantitative data items and ~10,000 words.Thematic analysis of the supporting reasons for scores is produced by training an NLP model to map reasons into a predetermined set of themes typically used in evaluating investments. Feedback from the evaluating teams and startup founding teams allow for continuous training of explanation system.We show that an archi- tecture of interoperable models are highly effective in achieving both accuracy and explanatory power in investment evaluations.
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 | 0.007 | 0.033 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.008 | 0.006 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.003 | 0.003 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.003 | 0.001 |
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".