Deploying AI4citations: From research to production
Building a production-ready citation verification system with automated testing, comprehensive APIs, and scalable deployment
Posts by Tag
- Machine Learning 5
- Data 2
- Python 1
- Feature Engineering 1
- Genomics 1
- Geochemistry 1
- R 1
- LLMs 1
- Transfer Learning 1
- TensorFlow 1
- Deep Learning 1
- NLP 1
- Claim Verification 1
- Overfitting 1
- Bias-Variance Tradeoff 1
- AI Apps 1
- User Feedback 1
- Concept Drift 1
- LLM 1
- RAG 1
- LangChain 1
- LangGraph 1
- ChromaDB 1
- BM25 1
Machine Learning
Modern understanding of overfitting and generalization in machine learning
In a previous post, I described my experience with fine-tuning transformer models on small NLP datasets with less than a few thousand examples. One of the go...
ML MiniLesson: Using pretrained layers
Transfer learning take layers of an existing neural network previously trained on a (usually large) dataset. These layers are the starting point to build a n...
Sustainable data preprocessing with pipelines
Early in my modeling of the Titanic dataset (a kind of “Hello World” for machine learning), I was struck by the variety - or inconsistency - of data preproce...
Installing AutoSklearn with conda
I really wanted to try out the examples of using AutoSklearn over at Machine Learning Mastery. But when I ran pip install auto-sklearn, an error occurred:
Data
Thermodynamic data consistency: The case of Li-mica
How do ore deposits form? This question motivates geochemists to build internally consistent thermodynamic databases to predict conditions where minerals pre...
Sustainable data preprocessing with pipelines
Early in my modeling of the Titanic dataset (a kind of “Hello World” for machine learning), I was struck by the variety - or inconsistency - of data preproce...
Python
Installing AutoSklearn with conda
I really wanted to try out the examples of using AutoSklearn over at Machine Learning Mastery. But when I ran pip install auto-sklearn, an error occurred:
Feature Engineering
Hypothesis-driven feature engineering
I’m starting a machine learning project to use microbial communities as biosensors for redox potential. Why is this important? Well, redox potential (written...
Genomics
Hypothesis-driven feature engineering
I’m starting a machine learning project to use microbial communities as biosensors for redox potential. Why is this important? Well, redox potential (written...
Geochemistry
Thermodynamic data consistency: The case of Li-mica
How do ore deposits form? This question motivates geochemists to build internally consistent thermodynamic databases to predict conditions where minerals pre...
R
Thermodynamic data consistency: The case of Li-mica
How do ore deposits form? This question motivates geochemists to build internally consistent thermodynamic databases to predict conditions where minerals pre...
LLMs
LLMs in claim verification
Scientific papers make lots of claims with supporting citations. However, even well-intentioned writers make mistakes, and automated citation verification is...
Transfer Learning
ML MiniLesson: Using pretrained layers
Transfer learning take layers of an existing neural network previously trained on a (usually large) dataset. These layers are the starting point to build a n...
TensorFlow
ML MiniLesson: Using pretrained layers
Transfer learning take layers of an existing neural network previously trained on a (usually large) dataset. These layers are the starting point to build a n...
Deep Learning
Experimenting with transformer models for citation verification
Automated verification of scientific claims represents a critical application of natural language processing in maintaining research integrity. The exponenti...
NLP
Experimenting with transformer models for citation verification
Automated verification of scientific claims represents a critical application of natural language processing in maintaining research integrity. The exponenti...
Claim Verification
Experimenting with transformer models for citation verification
Automated verification of scientific claims represents a critical application of natural language processing in maintaining research integrity. The exponenti...
Overfitting
Modern understanding of overfitting and generalization in machine learning
In a previous post, I described my experience with fine-tuning transformer models on small NLP datasets with less than a few thousand examples. One of the go...
Bias-Variance Tradeoff
Modern understanding of overfitting and generalization in machine learning
In a previous post, I described my experience with fine-tuning transformer models on small NLP datasets with less than a few thousand examples. One of the go...
AI Apps
Deploying AI4citations: From research to production
Building a production-ready citation verification system with automated testing, comprehensive APIs, and scalable deployment
User Feedback
Deploying AI4citations: From research to production
Building a production-ready citation verification system with automated testing, comprehensive APIs, and scalable deployment
Concept Drift
Combating concept drift in LLM applications
Your LLM application works beautifully at launch. The model understands user preferences, accuracy metrics look great, and feedback is positive. Then, gradua...
LLM
Combating concept drift in LLM applications
Your LLM application works beautifully at launch. The model understands user preferences, accuracy metrics look great, and feedback is positive. Then, gradua...
RAG
Engineering challenges for a real RAG application
A case study with R-help-chat
LangChain
Engineering challenges for a real RAG application
A case study with R-help-chat
LangGraph
Engineering challenges for a real RAG application
A case study with R-help-chat
ChromaDB
Engineering challenges for a real RAG application
A case study with R-help-chat
BM25
Engineering challenges for a real RAG application
A case study with R-help-chat