Skip to content

Instantly share code, notes, and snippets.

@computerphysicslab
Created September 25, 2020 12:26
Show Gist options
  • Star 0 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save computerphysicslab/7e81131e94d537ffec01c2b3d3de4570 to your computer and use it in GitHub Desktop.
Save computerphysicslab/7e81131e94d537ffec01c2b3d3de4570 to your computer and use it in GitHub Desktop.
Idea/BackLog/Proyecto de Notas + AI:
Agregador online de recortes
Clasificador automático
#quickAnswers #gitGists #knowledgeManagement #KM #collectiveKnowledge #collectiveIntelligence #wisdomOfCrowd
Knowledge management (KM) is the process of creating, sharing, using and managing the knowledge and information of an organization.
Ejemplo 1:
“Featured Snippets” are being driven by a powerful AI/Machine Learning algorithms called “sentence-compression algorithms”. Able to determine the most relevant sentence or paragraph within a page and extract it programmatically to popular the featured snippet box.
https://alamedaim.com/google-featured-snippets/
Es un recorte de Internet sobre un tema interesante que se convierte en una nota compartida, para referencia futura.
La autoclasificación la colocaría en la carpeta AI
La expansión semántica le añadiría los hashtags: #AI #machineLearning #semanticAnalysis #textSummarization
Ejemplo 2:
"Abstractive summarization methods w/ deep learning. These models consist of an encoder and a decoder, where a neural network reads the text, encodes it and then generates target text."
"First, extractive summarization systems form summaries by copying and rearranging passages from the original text. Second, abstractive summarization systems generate new phrases, rephrasing or using words that were not in the original text. Due to the difficulty of abstractive summarization, the great majority of past work has been extractive."
https://medium.com/luisfredgs/automatic-text-summarization-with-machine-learning-an-overview-68ded5717a25
#AI #NLP #textSummarization #sentenceRanking #wordVectors #extractiveSummarization #latentSemanticAnalysis #LSA
UX:
Un bookmarklet que:
1.- recoge el texto seleccionado y la URL
2.- generando un snippet
3.- almacenándolo en el cloud
4.- organizándolo en categorías
5.- expansión semántica, topic suggestion
6.- agregándolo al workgroup (Jira/Trello)
7.- público + comentarios (yCombinator)
Analizar si esto ya existe...
+ Trello es Kanban cards muy sencillo
+ Jira es el Kanban cards más formal
+ clipboard bookmarklet: https://gist.github.com/bradleybossard/3667ad5259045f839adc No funciona
+ page load stats bookmarklet: https://github.com/micmro/performance-bookmarklet
+ Automatically Categorized Bookmarks: https://stash.ai/ $36/year - public & private
+ Tagpacker
+ Raindrop.io - All your articles, photos, video & content from web & apps in one place.
+ Pocket - When you find something you want to view later, put it in Pocket.
+ Diigo - Diigo is a powerful research tool and a knowledge-sharing community
+ Pinboard - Pinboard is a personal archive for things you find online and don't want to forget.
+ Papaly - Papaly is a simple social bookmarking tool to manage all your links.
+ Google Bookmarks - With Google Bookmarks, you can save shortcuts to your favorite webpages and navigate to them in seconds from anywhere.
Ideas y Contexto:
Estas funcionalidades son aplicables a otros sistemas de KM:
+ email: idealmente organizado por carpetas (automáticas?) y búsqueda semántica potente
+ googlePhotos: estaría bien que autoordenara por carpetas y pusiera hashtags automáticos
+ incidencias: un sistema de incidencias, incluso con triaje (determinación de prioridad y categoría) automático
+ git: repositorios de código, con vistas autoclasificadas, tags optimizados y sugerencias para el backlog
+ healthcare: historia clínica:
* Para el médico: Vista de pacientes por categorías y exploración de casos más similares
* Para investigación: tras anonimización:
- detección de outsiders (errores o bien enfermedades extrañas)
- training de sistemas de aprendizaje ML (casos parecidos => diagnósticos frecuentes)
+ Posts en redes sociales
+ crawler: snippet, keywords, class, similar, score?
El punto en común entre todos ellos es que:
1.- Se recibe un nuevo elemento (bookmarklet)
2.- Se extraen las features
3.- AI: training + features =>
* Se cataloga: clustering
* Se añaden hashtags: curation
* Encuentra elementos similares rankeados: recommender system
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment