PUMA Vault
Search
Buscar
Modo oscuro
Modo claro
Explorador
Home
❯
30 Permanent
❯
31 Concepts
Carpeta: 30---Permanent/31-Concepts
60 artículos en esta carpeta.
15 jun 2026
LLM-Concepts
15 jun 2026
AI systems can generate new scientific knowledge, but only within human-defined research frameworks
permanent-note
ai-science
knowledge-generation
puma
agentic-science
academic-writing
accuracy
ai-scientist
alphafold
anthropic
code-review
critical-thinking
gemini
github
gnome
google
gpt
graphcast
hypothesis
llm
materials-science
metrics
moc
multi-agent
notebooklm
openai
pipeline
plasma-physics
project-management
protein-folding
red-teaming
reinforcement-learning
research
research-methodology
research-tools
scientific-knowledge
smart-pmo
weather-prediction
15 jun 2026
PN — PUMA's PM problem framing rests on three TIC governance frameworks treated…
permanent-note
puma
reference
agile
governance
15 jun 2026
PN — Flyvbjerg & Gardner (2023) introduce the 'uniqueness trap'
permanent-note
puma
reference
15 jun 2026
PN — PUMA's datasets derive from two verified benchmarks: the Jira Social…
permanent-note
puma
reference
story-points
dataset
tawos
jira
15 jun 2026
PN — The reproducibility contribution is grounded on Angermeir et al
permanent-note
puma
reference
reproducibility
15 jun 2026
PN — PUMA's prompt strategies are anchored to seminal prompting literature:…
permanent-note
puma
reference
prompting
few-shot
chain-of-thought
zero-shot
15 jun 2026
PN — Story points in PUMA follow the canonical Fibonacci series {1,2,3,5,8,13,21}
permanent-note
puma
reference
story-points
tawos
15 jun 2026
PN — RCOIF is PUMA's own prompt-engineering framework
permanent-note
puma
reference
rcoif
15 jun 2026
Active reading turns passive information into actionable knowledge through deliberate questioning
permanent-note
reading
cognition
research-methodology
mit-ai-lab
keshav
academic-writing
ai-tools
anthropic
elicit
gtd
literature-review
moc
notebooklm
para
perplexity
pipeline
reading-method
research
research-tools
slr
wp316
zettelkasten
15 jun 2026
Agentic Science represents a new research paradigm where AI manages complete investigation cycles autonomously
permanent-note
agentic-science
robot-scientists
ai-discovery
puma
academic-writing
ai-science
ai-scientist
alphafold
automation
code-review
github
gnome
hypothesis
issue-triage
llm
materials-science
moc
multi-agent
orchestration
planning
project-management
protein-folding
react
reasoning
reasoning-action
research
research-methodology
smart-pmo
sprint
triage
15 jun 2026
Algorithmic Bias in AI-Assisted Project Management
permanent
algorithmic-bias
fairness
ethics
discrimination
project-management
ai-safety
puma-core
research
hitl
issue-triage
effort-estimation
accountability
social-impact
diversity
15 jun 2026
Prompting Frameworks — CO-STAR, Self-Consistency, and Structured Prompting
permanent
prompting
costar
self-consistency
chain-of-thought
few-shot
zero-shot
structured-output
prompt-engineering
puma-core
research
agents
llm
benchmark
issue-triage
effort-estimation
architecture
15 jun 2026
Computational Sustainability — Carbon Footprint of LLM Experiments
permanent
sustainability
carbon-footprint
co2
codecarbon
green-ai
energy
emissions
puma-core
research
ethics
llm
benchmark
local-models
experiment
cloud-vs-local
15 jun 2026
Evaluation Metrics — Comprehensive Reference for PUMA
permanent
evaluation
metrics
f1
mae
rmse
spearman
wilcoxon
auc-roc
bleu
rouge
perplexity
co2
inference-latency
puma-core
research
benchmark
statistics
story-points
issue-triage
effort-estimation
statistical-validation
nlp
sustainability
15 jun 2026
Fine-Tuning LLMs — LoRA, QLoRA, GGUF Quantization, and PUMA Considerations
permanent
fine-tuning
lora
qlora
quantization
gguf
ollama
llm
training
puma-core
research
agents
benchmark
local-models
effort-estimation
issue-triage
architecture
15 jun 2026
Generative Agents — Memory Stream, Reflection, and Planning Architecture
permanent
generative-agents
memory-stream
reflection
planning
agent-architecture
llm
simulation
emergent-behavior
puma-core
research
agents
architecture
multi-agent
smart-pmo
persistent-memory
15 jun 2026
Human-in-the-Loop (HITL) and Bounded Autonomy for AI Agents
permanent
hitl
human-in-the-loop
bounded-autonomy
ai-safety
ethics
oversight
control
human-ai-collaboration
puma-core
research
agents
architecture
project-management
accountability
alignment
15 jun 2026
Local LLM inference trades capability ceiling for reproducibility, privacy, and zero marginal cost
permanent-note
llm
local-inference
reproducibility
sustainability
puma
anthropic
api
benchmark
carbon-footprint
code-review
critical-thinking
effort-estimation
embeddings
falsifiability
github
gpt
llama
local-llm
meta
mistral
moc
ollama
openai
popper
privacy
project-management
rag
reasoning
red-teaming
retrieval
software-engineering
story-points
vector-db
15 jun 2026
LLM Models Used in PUMA — Technical Reference
permanent
llm
models
llama
mistral
phi
gemma
deepseek
gpt4o
qwen
moe
quantization
ollama
puma-core
research
benchmark
effort-estimation
issue-triage
agents
architecture
15 jun 2026
PN — The project asks whether local language models can manage technology…
permanent-note
puma
methodology
15 jun 2026
PN — PUMA is a recursive acronym ('PUMA Understanding & Management w Agents')
permanent-note
puma
methodology
15 jun 2026
PN — A historical precedent (an early automated content bot that propagated a…
permanent-note
puma
methodology
dataset
15 jun 2026
PN — Technology project management in multi-project organizations suffers a…
permanent-note
puma
methodology
15 jun 2026
PN — Problem one
permanent-note
puma
methodology
triage
dataset
bias
15 jun 2026
PN — Problem two
permanent-note
puma
methodology
estimation
dataset
agile
15 jun 2026
PN — Problem three
permanent-note
puma
methodology
15 jun 2026
PN — Reproducibility gap: in the most complete field review (85 papers across…
permanent-note
puma
methodology
reproducibility
15 jun 2026
PN — Gap two
permanent-note
puma
methodology
estimation
prompting
15 jun 2026
PN — Gap three
permanent-note
puma
methodology
sustainability
carbon
metrics
15 jun 2026
PN — PUMA occupies a deliberately narrow niche
permanent-note
puma
methodology
carbon
benchmark
local-llm
15 jun 2026
PN — Meeting the estimation MVP threshold and the H2 hypothesis contrast are…
permanent-note
puma
methodology
estimation
hypothesis
15 jun 2026
PN — A transferable lesson: on a long-tailed story-point distribution, the…
permanent-note
puma
methodology
mae
baseline
15 jun 2026
PN — The title question is answered cautiously
permanent-note
puma
methodology
triage
estimation
baseline
15 jun 2026
PN — The H2 non-rejection is reframed as a valuable methodological finding…
permanent-note
puma
methodology
15 jun 2026
PN — The environmental framing
permanent-note
puma
methodology
15 jun 2026
PN — The diversity/inclusion stance is open-source MIT release executable…
permanent-note
puma
methodology
release
15 jun 2026
Multi-agent systems outperform single agents on PM tasks when agent roles match task specialisation boundaries
permanent-note
multi-agent
architecture
specialisation
pm-agents
masai
metagpt
artefact
backlog
baseline
benchmark
bmad
critical-thinking
dsr
effort-estimation
evaluation
gpt
human-in-the-loop
hypothesis
issue-triage
llm
moc
openai
orchestration
planning
project-management
react
reasoning
reasoning-action
red-teaming
research-methodology
sdd
smart-pmo
spec-driven-development
story-points
swe-bench
triage
15 jun 2026
PN — PUMA is a local-first benchmarking platform that evaluates open-weight…
permanent-note
puma
engine
ollama
benchmark
15 jun 2026
PN — H1 (triage): qwen2.5:3b + contextual-anchoring (F1 0.5894) statistically…
permanent-note
puma
engine
triage
wilcoxon
contextual-anchoring
baseline
15 jun 2026
PN — H2 (estimation): best config qwen2.5:7b cuts MAE only 0.17 SP vs a…
permanent-note
puma
engine
estimation
mae
baseline
15 jun 2026
PN — Model scaling is non-monotonic on these PMO tasks: in 4 of 6 (family x…
permanent-note
puma
engine
triage
local-llm
15 jun 2026
PN — PUMA's sustainability measurement is explicitly bounded: it covers the…
permanent-note
puma
engine
sustainability
15 jun 2026
PN — The +/-0.01 baseline tolerance does not absorb seed variance (zero at…
permanent-note
puma
engine
ollama
baseline
15 jun 2026
PUMA is an applied Agentic Science system positioned at LeCun's Level 2 of AI scientific capability
permanent-note
puma
agentic-science
positioning
lecun
academic-writing
ai-ethics
ai-science
ai-scientist
alphafold
anthropic
automation
baseline
code-review
ethics
evaluation
github
gnome
graphcast
hypothesis
issue-triage
jira
literature-review
llm
materials-science
moc
multi-agent
notebooklm
orchestration
project-management
protein-folding
research-methodology
research-tools
slr
smart-pmo
triage
weather-prediction
workflow
15 jun 2026
RAG, embeddings, and vector databases enable LLMs to query external knowledge without fine-tuning
permanent-note
rag
embeddings
vector-database
llm
puma
architecture
chain-of-thought
cot
critical-thinking
dataset
effort-estimation
few-shot
fine-tuning
issue-triage
jira
llama
local-llm
memory
meta
metrics
moc
nlp
ollama
pipeline
precision-recall
project-management
react
reasoning-action
red-teaming
retrieval
story-points
triage
vector-db
15 jun 2026
The ReAct pattern reduces LLM hallucination by grounding each reasoning step in observable external evidence
permanent-note
react
hallucination
grounding
agent-pattern
reasoning
api
chain-of-thought
cot
effort-estimation
embeddings
few-shot
hypothesis
issue-triage
llm
metrics
moc
multi-agent
precision-recall
project-management
rag
reasoning-action
research-methodology
retrieval
story-points
triage
vector-db
zettelkasten
15 jun 2026
Reflexion — Verbal Self-Reflection for Agent Self-Improvement
permanent
reflexion
self-reflection
agent
llm
verbal-reinforcement
self-critique
episodic-memory
actor-evaluator
puma-core
research
agents
architecture
benchmark
coding
issue-triage
effort-estimation
iterative-improvement
15 jun 2026
Statistical Validation — Wilcoxon, Bootstrap, Effect Size, and CI for PUMA
permanent
statistics
wilcoxon
bootstrap
effect-size
confidence-interval
shapiro-wilk
cohen-d
non-parametric
hypothesis-testing
puma-core
research
evaluation
benchmark
effort-estimation
story-points
statistical-validation
15 jun 2026
PUMA tool selection follows a hierarchy: reproducibility first, then capability, then convenience
permanent-note
tool-selection
reproducibility
local-inference
puma-design
api
architecture
autogen
code-review
effort-estimation
embeddings
github
gpt
langgraph
llama
llm
lm-studio
local-llm
meta
metrics
mistral
moc
ollama
openai
precision-recall
pydantic
python
rag
retrieval
story-points
vector-db
15 jun 2026
The Uniqueness Trap — Why PM Estimation Fails and How PUMA Fixes It
permanent
uniqueness-trap
flyvbjerg
reference-class-forecasting
planning-fallacy
cognitive-bias
project-management
puma-core
research
effort-estimation
story-points
benchmark
critical-thinking
scope-creep
cost-overrun
01 may 2026
Context Engineering — Designing the LLM Context Window as a System
permanent
context-engineering
prompt-engineering
llm
context-window
rag
memory
tools
system-prompt
puma-core
research
agents
architecture
prompting
planning
agent-design
information-retrieval
01 may 2026
MCP — Model Context Protocol: Architecture, Security, and PUMA Integration
permanent
mcp
model-context-protocol
anthropic
tool-use
agent-infrastructure
jira
github
smart-pmo
security
puma-core
research
api
architecture
agentic-ai
protocol
standards
ethics
governance
01 may 2026
RLHF and Constitutional AI — LLM Alignment Training Paradigms
permanent
rlhf
constitutional-ai
rlaif
ppo
reward-model
alignment
ai-safety
fine-tuning
sft
anthropic
openai
puma-core
research
training
ethics
llm
models
hitl
01 may 2026
Transformer Architecture and Mixture of Experts — Technical Reference for PUMA Models
permanent
transformer
attention
moe
mixture-of-experts
architecture
deepseek
mixtral
llama
scaling
puma-core
research
llm
models
technical-reference
self-attention
feed-forward
positional-encoding
01 may 2026
Tree of Thoughts — Deliberate Multi-Path Reasoning for LLMs
permanent
tree-of-thoughts
tot
deliberate-reasoning
planning
backtracking
bfs
dfs
beam-search
cot
llm-reasoning
puma-core
research
prompting
agents
reasoning
issue-triage
backlog-prioritisation
01 mar 2026
Chain-of-Thought (CoT) Prompting
permanent
concept
prompting
cot
reasoning
llm
academic-writing
api
baseline
bias
chain-of-thought
dashboard
dataset
effort-estimation
ethics
evaluation
few-shot
hypothesis
issue-triage
jira
llama
meta
metrics
mistral
moc
permanent-note
precision-recall
project-management
rcoif
react
reasoning-action
research
research-methodology
story-points
tawos
template
triage
zero-shot
zettelkasten
01 mar 2026
Issue Triage in Software Project Management
permanent
concept
triage
project-management
jira
priority
classification
academic-writing
agile
automation
baseline
bias
chain-of-thought
code-review
cot
critical-thinking
dataset
effort-estimation
ethics
evaluation
few-shot
github
glossary
gpt
hypothesis
ict
issue-triage
llm
metrics
moc
openai
permanent-note
planning
precision-recall
red-teaming
research
research-methodology
security
software-engineering
sprint
story-points
tawos
01 mar 2026
LLM Agents — Definition and Taxonomy
permanent
concept
llm-agents
agents
orchestration
ai-ethics
ami
anthropic
api
architecture
artefact
baseline
bias
chain-of-thought
code-review
cot
crewai
critical-thinking
data-formats
dataset
dev-tools
devops
docker
drca
dsr
effort-estimation
ethics
evaluation
few-shot
github
hypothesis
iipr
issue-triage
jira
json
langgraph
llama
llm
local-llm
memory
meta
mistral
moc
multi-agent
ollama
opencode
permanent-note
planning
project-management
prompt-engineering
prompting
python
rag
rcoif
reasoning
red-teaming
research-methodology
retrieval
software-engineering
story-points
tawos
tool-use
triage
validity
zero-shot
zettelkasten
01 mar 2026
🧠 Concepts — Zettelkasten Guide
readme
concepts
zettelkasten
permanent
agile
backlog
chain-of-thought
cot
few-shot
issue-triage
llm
project-management
rag
retrieval
sprint
template
triage