amtech / ai-tensor-house

A collection of reference machine learning and optimization models for enterprise operations: marketing, pricing, supply chain

About

TensorHouse is a collection of reference machine learning and optimization models for enterprise operations: marketing, pricing, supply chain, and more. The goal of the project is to provide baseline implementations for industrial, research, and educational purposes.

The project focuses on models, techniques, and datasets that were originally developed either by industry practitioners or by academic researchers who worked in collaboration with leading companies in technology, retail, manufacturing, and other sectors. In other words, TensorHouse focuses mainly on industry-proven methods and models rather than on theoretical research.

TensorHouse contains the following resources:

• a well-documented repository of reference notebooks and applications (templates),
• a manually curated list of important papers in modern operations research,
• a manually curated list of public datasets related to enterprise use cases.

Illustrative Examples

Strategic price optimization using reinforcement learning: DQN learns a Hi-Lo pricing policy that switches between regular and discounted prices

Supply chain optimization using reinforcement learning: Diagnostic interface for the simulation environment

Anomaly detection in images using autoencoders: Anomaly masks for defect location detection

List of Models

Promotions, Offers, and Advertisements

• Media Mix, Attribution, and Budget Optimization
  • Media Mix Modeling: Basic Adstock Model for Campaign/Channel Attribution (notebook)
  • Media Mix Modeling: Bayesian Model with Carryover and Saturation Effects (notebook)
  • Multitouch Channel Attribution Model Using Deep Learning (LSTM with Attention) (notebook)
• Customer Scoring and Lifetime Value
  • Promotion Effect Estimation Using Causal Inference Methods (Regression and Mathing) (notebook)
  • Customer Propensity Scoring Using Deep Learning (LSTM with Attention) (notebook)
  • Customer Lifetime Value (LTV) Modeling Using Markov Chain (notebook)
• Decision Automation
  • Dynamic Content Personalization Using Contextual Bandits (LinUCB) (notebook)
  • Next Best Action Model Using Reinforcement Learning (Fitted Q Iteration) (notebook)

Search

• Text Search
  • Latent Semantic Analysis (LSA) (notebook)
  • Retrieval-augmented Generation (RAG) Using LLMs (notebook)
  • Retrieval-augmented Generation (RAG) Using LLMs Agents (notebook)
• Visual Search
  • Visual Search by Artistic Style (VGG16) (notebook)
  • Visual Search based on Product Type (EfficientNetB0) (notebook)
  • Visual Search Using Variational Autoencoders (notebook)
  • Image Search Using a Language-Image Model (CLIP) (notebook)
• Structured Data Search
  • Relational Data Querying Using LLMs (notebook)
• Data Preprocessing
  • Product Attribute Discovery, Extraction, and Harmonization Using LLMs (notebook)

Recommendations

• Embedding Calculation
  • Item2Vec Model Using Word2vec (notebook)
  • Customer2Vec Model Using Doc2vec (notebook)
• Collaborative Filtering
  • Nearest Neighbor User-based Collaborative Filtering (notebook)
  • Nearest Neighbor Item-based Collaborative Filtering (notebook)
• Deep and Hybrid Recommenders
  • Neural Collaborative Filtering - Prototype (notebook)
  • Neural Collaborative Filtering - Hybrid Recommender (notebook)
  • Behavior Sequence Transformer (notebook)
  • Graph Recommender Using Node2Vec (notebook)

Content Analytics

• Sentiment Analysis (notebook)

Demand Forecasting

• Traditional Methods
  • Demand Forecasting Using Exponential Smoothing (ETS) (notebook)
  • Demand Forecasting and Price Elasticity Analysis Using Time Series Regression (notebook)
• Deep Learning Methods
  • Demand Forecasting Using DeepAR (notebook)
  • Demand Forecasting Using NeuralProphet (notebook)
• Data Preprocessing
  • Demand Unconstraining (notebook)

Pricing and Assortment

• Static Price, Promotion, and Markdown Optimization
  • Market Response Functions (notebook)
  • Price Optimization for Multiple Products (notebook)
  • Price Optimization for Multiple Time Intervals (notebook)
• Dynamic Pricing
  • Dynamic Pricing Using Thompson Sampling (notebook)
  • Dynamic Pricing with Limited Price Experimentation (notebook)
  • Bayesian Demand Models (notebook)
  • Price Optimization Using Reinforcement Learning (DQN) (notebook)

Supply Chain

• Single-echelon Inventory Optimization Using (s,Q) and (R,S) Policies (notebook)
• Multi-echelon Inventory Optimization Using Reinforcement Learning (DDPG, TD3) (notebook)
• Inventory Allocation Optimization (notebook)
• Supply Chain Simulator for Reinforcement Learning Based Optimization (PPO) (notebook)
• Supply Chain Control Tower Using LLMs (notebook)

Anomaly Detection

• Noise Reduction in Multivariate Timer Series Using Linear Autoencoder (PCA) (notebook)
• Remaining Useful Life Prediction Using Convolution Networks (notebook)
• Anomaly Detection in Time Series (notebook)
• Anomaly Detection in Images using Autoencoders (notebook)

Basic Templates

• Generic Regression and Classification Models

  • Neural Network with Vector Inputs (notebook)
  • Neural Network with Sequential Inputs (ConvNet, LSTM, Attention) (notebook)

• Enterprise Time Series Analysis

  • Forecasting Using ARIMA and SARIMA (notebooks 1 2)
  • Decomposition and Forecasting using Bayesian Structural Time Series (BSTS) (notebooks 1 2 3 4)
  • Forecasting and Decomposition using Gradient Boosted Decision Trees (GBDT) (notebook)
  • Forecasting and Decomposition using LSTM with Attention (notebook)
  • Forecasting and Decomposition using VAR/VEC models (notebooks 1 2)

Approach

• The most basic models come from the Introduction to Algorithmic Marketing book.
  • Book's website - https://www.algorithmicmarketingbook.com/
• More advanced models use deep learning and reinforcement learning techniques from The Theory and Practice of Enterprise AI book.
  • Book's website - https://www.enterprise-ai-book.com/
• Most models are based on industrial reports and real-life case studies

Community

Follow our Twitter feed for notifications about new developments.