"No one is harder on a talented person than the person themselves" - Linda Wilkinson ; "Trust your guts and don't follow the herd" ; "Validate direction not destination" ;

February 26, 2023

NLP - NER - Entity Recognition

I have worked on NLP, and custom NER examples. This streamlit demo covers NER in domain context and multiple entities link






  • Name Entity Recognition - Extract Organizations, People, Locations, and many other entities from long, free-text financial documents.
  • Extract Financial Relationships - Automatically identify relationships between companies, products, and people – even when they are mentioned using aliases.
  • Classify Financial Text - Classify texts into 77 banking-related categories like credit reports, mortgages, money transfers and more.
  • Financial Sentiment Analysis - Identify positive, negative or neutral sentiments in financial news.
  • Financial De-identification - De-identify and mask sensitive personal information in documents and images.

Ref - Link

Keep Exploring!!!

runwayml - Image Experiments - Variations, Text to Image Generation

Looks great better than my segmentation efforts. A good start :)

App - Link


Background Mixing


Image Variations

More Tools


Keep Exploring!!!

Exploring landing.ai

Custom Detection model - Low Code / No Code Platform

  • Reloading - Link
  • Labeling custom forest detection data
  • Training
  • Predicting
Labeling

Training


Deploying


Predicting



Keep Exploring!!!

February 25, 2023

Full Stack vs Deep Stack

I always feel myself an aspirational Deep Stack guy, Getting better in one focused area and expanding on related areas. I don't think it would be right to stay expert in all vs exposure in all.

  • Expertise vs Exposure
  • Communication vs Capability
  • Consistency vs Competency

Everything will be reflected in our plans, actions, and thinking. 

Ref - Link

In my career, I prefer to be a good T in some areas and Try to be a V where to build solutions I need to learn. 70% T and 30% V, You only have limited time to keep sharpening skills vs catching up on related skills. 



As a Team, You need a mix of all ingredients


Ref - Link

Go where you Grow, Grow where you Go. 
Titles <> Knowledge
Keep Learning

Keep Thinking!!!

February 24, 2023

This week in AI - ML World - Feb 24 2023

This week some key observations

  • Hugging Face + AWS Partnership - 10K plus Hugging Face + AWS Ecosystem more options for customers
  • Coca-Cola Signs As Early Partner for OpenAI’s ChatGPT, DALL-E Generative AI - Expect more tags, oneliners, new images, and eye-catching images.
  • Bring your data, Build your LLM model seems to be another business offering from OpenAI
  • With GAN models, everything converges. The problem is already solved with  LargeLanguageModels. Evolve to the next level or perish. You may have L1 / L2 already in your product, You may be onboarding L3 to finetune better solutions
  • Nvidia Fintech Survey 2023, Key use cases of focus this year.

How to handle LLM, Threats, or collaboration Opportunities?
Continue to collect more data, Finuetune in-house models, Leverage In-house, Finetune, LLM and have voting on top of it to pick based on best-aligned response. 

Three Phased Approach

  • Custom NER, Parsing, Intent Extraction (Level 1 - Inhouse)
  • Leverage Other lightweight fine-tuned models for summarization, and  topics (Level 2 - Finetuned)
  • Leverage LLM models to rank/ Evaluate results (Level 3 - LLM Inputs for finetuning)

Interesting Demos



Currently, there is a lot of noise with GPT, funding. Few will survive. Now the news about crypto, web3, metaverse, and bitcoin seems to have given up for GPT

Keep Exploring!!!

How To Enable GUI On AWS EC2 Ubuntu server

 A very needed read for me on enabling GUI in EC2 instances

How To Enable GUI On AWS EC2 Ubuntu server




Keep Exploring!!!

ML - Model Shipping Factories

5 years ago, #startups were in areas/segments 

  • AI-Driven Sales -Forecast, Recommendations (Data)
  • Chatbots - (NLP, Data)
  • Autopilot ADAS - Vision - Image, Video, Data
  • BPO - Customer support - NLP, Data, OCR
Companies that manufacture large models (Ref)


Vision / Image



Text




Audio



With GAN models, everything converges. The problem is already solved with LargeLanguageModels. Evolve to the next level or perish. #GAN #AI #startups. 

Keep Exploring!!!

MLFlow on AWS EC2

  • MLflow is organized into four components: Tracking, Projects, Models, and Model Registry. 
  • Create AWS Free EC2 t2 micro ubuntu machine
  • Follow the below steps to setup mlflow
  • Install ngnix to route requests from external networks
  • Run a few experiments, Visualize results
Steps


References





Keep Exploring!!!


February 23, 2023

Disruptive, Fast Forward GPT Days

News #1 - Bain & OpenAI Lineup to Solve more cases

Key use cases they target with ChatGPT are

  • Building next-generation contact centers for retail banks, telco and utility companies to support sales and service agents with automated, personalized, and real-time scripts, and to improve customer experience.
  • Boosting turn-around time for leading product and service marketers by using ChatGPT and DALL·E to develop highly personalized ad copy, rich imagery, and targeted messaging.
  • Helping financial advisors improve their productivity and responsiveness to clients through the analysis of client dialogues and financial literature, and the generation of digital communication.

News #2 - AWS and Hugging Face Team up for more adoption, focused solutions

Hugging Face has become the central hub for machine learning, with more than 100,000 free and accessible machine learning models. More solutions with AWS Platform

News #3 - You can also leverage ChatGPT to build your own model

OpenAI’s Foundry will let customers buy dedicated compute to run its AI models

The cost will be expensive though - Instances won’t be cheap. Running a lightweight version of GPT-3.5 will cost $78,000 for a three-month commitment or $264,000 over a one-year commitment. To put that into perspective, one of Nvidia’s recent-gen supercomputers, the DGX Station, runs $149,000 per unit.

News #4 - Coca-Cola Signs As Early Partner for OpenAI’s ChatGPT, DALL-E Generative AI

Coca-Cola will team with OpenAI and Bain & Company to use OpenAI’s ChatGPT and DALL-E platforms to craft personalized ad copy, images, and messaging, the companies announced in a press release. 

Expect more tags, oneliners, new images, and eye-catching images.

Three Phased Approach

  • Custom NER, Parsing, Intent Extraction (Level 1 - Inhouse)
  • Leverage Other lightweight fine-tuned models for summarization, topics (Level 2 - Finetuned)
  • Leverage LLM models to rank/ Evaluate results (Level 3 - LLM Inputs)

Keep Exploring!!!

Startup Analysis - hyperverge - KYV - Vision + OCR + NLP

Many times taking an idea, ideating it, and solving it end to end is key. KYC with Vision / Image / Data and NLP are very impressive.

Product Features

  • Real-time analysis of images and videos obtained from sources such as consumer photos, satellite images, surveillance cameras, industrial images, and documents. 
  • NLP solutions for automating and disrupting the Legal Document Analysis industry

Deep Learning Skills (Vision / NLP) - Our Perspectives

  • DL - Models built for Face detection, OCR, and Buildings / Signs Detection. Face, Object, Text, and Activity recognition.
  • NLP - One shot, Few shot, Self-Supervised approach, multi-task learning, and contrastive learning strategies
  • Plus a lot of custom embeddings/graph databases/custom models

Video KYC

  • Liveliness Check
  • Background Detection
  • Landmark validations
  • Key facial landmarks based on submitted docs
  • Similarity scores
  • Social media similar image scores 

Signature Check

  • Signature Font Size, Length, Height, features of it
  • Signature Font Style
  • Keypoints match, landmarks match, shape, texture

Face Verification

  • Landmarks
  • Landmark distances for iris, nose, cheek, chin
  • Mediapipe
  • Custom Segment and measure similarity
  • Classify face shapes/hairstyles

Keep Exploring!!!

February 22, 2023

Law and Order - Applied Vision, NLP, ML Use cases

Vision Use cases

  • Person Detection, Attribute Extraction
  • Real-time vehicle number plate validation
  • Attribute Extraction - Shirt, Dress Type
  • Age Estimation
  • Personal Re-identification
  • Vehicle Details Extraction
  • Vehicle color Extraction
  • Action Recognition - Detect Crowd, Fight
  • Anomaly Detection, Loitering - Notify suspicious movements

NLP Use cases

  • Using NLP to find similar cases from Digitized documents
  • Using OCR to digitize documents

ML Use cases

  • Cluster patterns of crimes
  • Cluster types of crime
  • Cluster patterns of offenders
  • Churn time for offenders
  • Average time after a conviction for repeat crimes

CCTV analysis is a reactive approach. Edge + AI + Data + NLP is the way forward. Be predictive, proactive, and prepared.

Analyze the below companies and build another product :)




Ref - Link, Link1

Mark43 Interesting Demo and Screenshots - Link




Crime Prediction using Machine Learning with a Novel Crime Dataset

  • News Link Collection using Manual Process
  • News Link Collection using Web Crawler
  • Filtering Crawled News Collection to Identify Crime News






Keep Exploring!!!