PUMA Vault
Search
Buscar
Modo oscuro
Modo claro
Explorador
Home
❯
30 Permanent
❯
32 Methods
Carpeta: 30---Permanent/32-Methods
47 artículos en esta carpeta.
01 abr 2026
Research-Methods
15 jun 2026
PN — F1-macro (unweighted mean of per-class F1) is PUMA's primary triage…
permanent-note
puma
reference
triage
f1-macro
15 jun 2026
PN — MAE (mean absolute error in story points) is the primary effort-estimation…
permanent-note
puma
reference
estimation
story-points
mae
15 jun 2026
PN — Bit-exact reproducibility is guaranteed in warm state via seed=42,…
permanent-note
puma
reference
reproducibility
ollama
15 jun 2026
PN — 'Veritas' is PUMA's six-principle protocol for generative-AI use in…
permanent-note
puma
reference
15 jun 2026
PN — The canonical parser enforces a three-level policy (strict match ->…
permanent-note
puma
reference
15 jun 2026
PN — Fixed inference parameters across all templates: temperature=0.0, seed=42,…
permanent-note
puma
reference
triage
estimation
15 jun 2026
PN — PUMA Community operates as a serverless voluntary federation
permanent-note
puma
reference
community
federation
15 jun 2026
PN — The trust model is explicit 'collaborative good faith': schema validation…
permanent-note
puma
reference
integrity
15 jun 2026
PN — A formal 30-entry risk register (six categories x five) governs PUMA Community
permanent-note
puma
reference
community
15 jun 2026
PN — 'Discovery-before-write': no document section or code is written without…
permanent-note
puma
reference
15 jun 2026
PN — Install and start with one command: clone github.com/pumacp/puma then run…
permanent-note
puma
reference
ollama
docker
15 jun 2026
PN — Prepare datasets before the first run with puma prepare-datasets, which…
permanent-note
puma
reference
dataset
tawos
jira
15 jun 2026
PN — Statistical workflow: puma wilcoxon --model-a A --model-b B --scenario S…
permanent-note
puma
reference
wilcoxon
bias
15 jun 2026
PN — Publish results: puma export-community --run-id <ID> produces a signed…
permanent-note
puma
reference
community
metrics
json-schema
15 jun 2026
PN — Backup discipline: the puma_data volume (SQLite DB of all results) is the…
permanent-note
puma
reference
sqlite
dataset
integrity
15 jun 2026
PN — SQLite maintenance SOPs: integrity via PRAGMA integrity_check, space…
permanent-note
puma
reference
sqlite
integrity
15 jun 2026
PN — PUMA Info traceability
permanent-note
puma
reference
15 jun 2026
PN — The maintenance reproducibility guarantee rests on fixed seed=42 and…
permanent-note
puma
reference
reproducibility
carbon
codecarbon
baseline
15 jun 2026
PN — The triage task uses a public dataset of 50,000+ real issues from…
permanent-note
puma
methodology
triage
dataset
15 jun 2026
PN — The estimation task uses a public dataset of 31,294 real user stories with…
permanent-note
puma
methodology
estimation
story-points
dataset
agile
15 jun 2026
PN — The main research question asks whether there are statistically…
permanent-note
puma
methodology
triage
estimation
prompting
dataset
15 jun 2026
PN — Hypotheses follow Popper's demarcation criterion: each is a refutable null…
permanent-note
puma
methodology
15 jun 2026
PN — H1 (triage): the null states no model/strategy configuration beats the…
permanent-note
puma
methodology
triage
f1-macro
wilcoxon
baseline
15 jun 2026
PN — H2 (estimation): the null states no example-enriched configuration beats…
permanent-note
puma
methodology
estimation
story-points
mae
wilcoxon
15 jun 2026
PN — The work adopts Design Science Research, operationalized through six…
permanent-note
puma
methodology
15 jun 2026
PN — Inference runs locally via a model server exposing an OpenAI-compatible…
permanent-note
puma
methodology
metrics
quantization
15 jun 2026
PN — The MVP's canonical model is a 3B-parameter quantized model (~1.9 GB),…
permanent-note
puma
methodology
triage
estimation
15 jun 2026
PN — The experiment is a reduced factorial design with two main independent…
permanent-note
puma
methodology
mae
reproducibility
sustainability
prompting
15 jun 2026
PN — Estimation maps user stories to a canonical Fibonacci story-point value…
permanent-note
puma
methodology
estimation
mae
media
15 jun 2026
PN — Reproducibility uses five mechanisms: fixed seed=42 propagated everywhere,…
permanent-note
puma
methodology
reproducibility
15 jun 2026
PN — Four ethical risks are explicitly managed: role displacement…
permanent-note
puma
methodology
chain-of-thought
dataset
fairness
bias
15 jun 2026
PN — All model execution is delegated to a local Ollama daemon over HTTP at the…
permanent-note
puma
engine
ollama
15 jun 2026
PN — A benchmark run is fully spec-driven: a RunSpec YAML validated with…
permanent-note
puma
engine
benchmark
15 jun 2026
PN — Failed response parses are recorded as None, excluded from metric…
permanent-note
puma
engine
15 jun 2026
PN — A full Phase B sweep ran 9 models x 3 scenarios = 27 runs (2,700…
permanent-note
puma
engine
sweep
15 jun 2026
PN — PUMA reports a multi-dimensional metric suite
permanent-note
puma
engine
f1-macro
mae
calibration
latency
15 jun 2026
PN — Calibration metrics (ECE over 10 bins, MCE, Brier) are derived from…
permanent-note
puma
engine
calibration
ollama
metrics
15 jun 2026
PN — Sustainability is reported per run via CodeCarbon (CO2 kg, energy kWh,…
permanent-note
puma
engine
mae
sustainability
carbon
codecarbon
15 jun 2026
PN — PUMA treats inference compute cost as a first-class reported metric…
permanent-note
puma
engine
mae
carbon
15 jun 2026
PN — CodeCarbon runs in tracking_mode='machine' so whole-machine accounting…
permanent-note
puma
engine
carbon
codecarbon
architecture
15 jun 2026
PN — Submission integrity uses a SHA-256 predictions_summary_hash over a…
permanent-note
puma
engine
json-schema
integrity
15 jun 2026
PN — PUMA's security posture rests on local inference (no SaaS path/token to…
permanent-note
puma
engine
benchmark
integrity
15 jun 2026
PN — CI security tooling wires pip-audit (CVE scanning), bandit (SAST),…
permanent-note
puma
engine
01 mar 2026
Design Science Research (DSR)
permanent
method
dsr
research-paradigm
hevner
peffers
academic-writing
ai-tools
anthropic
architecture
artefact
bdd
benchmark
bias
carbon-footprint
code-review
codecarbon
cornell-notes
dataset
effort-estimation
elicit
ethics
finer
github
ict
issue-triage
jira
literature-review
llm
local-llm
metrics
moc
non-parametric
note-taking
ollama
permanent-note
perplexity
pipeline
precision-recall
prisma
project-management
quantitative-research
research
research-methodology
research-tools
sdd
semantic-scholar
slr
software-engineering
spec-driven-development
statistics
story-points
sustainability
tawos
triage
validity
wilcoxon
workflow
01 mar 2026
Methods: Wilcoxon · FINER · Cornell
permanent
method
wilcoxon
statistics
finer
cornell
research-quality
academic-writing
ai-tools
anthropic
baseline
benchmark
bias
carbon-footprint
chain-of-thought
code-review
codecarbon
contextual-anchoring
cornell-notes
cot
dsr
effect-size
egi
elicit
ethics
evaluation
fairness
falsifiability
github
human-in-the-loop
hypothesis
ict
iipr
issue-triage
literature-review
llm
metrics
non-parametric
note-taking
obsidian
permanent-note
perplexity
pipeline
popper
precision-recall
prisma
privacy
project-management
prompting
python
rcoif
reasoning
research
research-methodology
research-tools
slr
software-engineering
sustainability
template
transformer
triage
vault
zotero
01 mar 2026
📐 Methods — Guide
readme
methods
guide
cornell-notes
dataset
dsr
finer
literature-review
non-parametric
note-taking
prisma
research-methodology
slr
software-engineering
statistics
wilcoxon