Yuki Shizuya – Medium

Yuki Shizuya

Pinned

Yuki Shizuya
in
Intuition

Understanding L1 and L2 regularization with analytical and probabilistic views

Derive L1 and L2 regularization via analytical and probabilistic solution

May 25

Understanding L1 and L2 regularization with analytical and probabilistic views

May 25

Yuki Shizuya
in
Generative AI

Unlocking Mixture-of-Experts (MoE) LLM : Your MoE model can be embedding model for free

Mixture-of-experts (MoE) LLM can be used as an embedding model for free.

5d ago

Unlocking Mixture-of-Experts (MoE) LLM : Your MoE model can be embedding model for free

5d ago

Yuki Shizuya
in
Generative AI

Pixtral 12B is released : Novel SOTA Open-Source Multimodal Language Model

MistralAI team released Pixtral 12B, a novel open-source multimodal model that outperforms other open-source models. How does it works and…

Oct 23

Pixtral 12B is released : Novel SOTA Open-Source Multimodal Language Model

Oct 23

Yuki Shizuya
in
Towards AI

Vision Embedding Comparison for Image Similarity Search: EfficientNet vs.

What are the differences in embeddings among these models?

Oct 12

Vision Embedding Comparison for Image Similarity Search: EfficientNet vs.

Oct 12

Yuki Shizuya
in
Generative AI

Unlocking the Diffusion Model Power!

The introduction of a novel tuning-free style transfer method: InstantStyle. How can it work without additional fine-tuning, and how can it…

Sep 29

Unlocking the Diffusion Model Power!

Sep 29

Yuki Shizuya
in
AI Advances

Novel personalized image generation: InstantID — Theoretical Understanding and Applications

InstantID is a novel personalized image generation based on the diffusion model. This blog covers the theoretical understanding and how to…

Sep 25

Novel personalized image generation: InstantID — Theoretical Understanding and Applications

Sep 25

Yuki Shizuya
in
AI Advances

Introduction of Multimodal Topic Modeling

Learn how to implement multimodal topic modeling using CLIP and BLIP-2 with BERTopic

Sep 10

Introduction of Multimodal Topic Modeling

Sep 10

Yuki Shizuya
in
AI Advances

Explainable Clustering — The Introduction of Recursive Embedding and Clustering and its Application

Spotify developed the simple but powerful explainable clustering method

Aug 23

Explainable Clustering — The Introduction of Recursive Embedding and Clustering and its Application

Aug 23

Yuki Shizuya
in
Towards AI

Explainable AI for CLIP: The Architecture Explanation and its Application for Segment Anything

A detailed explanation of the effective activation map visualization technique for CLIP and its application for Segment Anything

Aug 13

Explainable AI for CLIP: The Architecture Explanation and its Application for Segment Anything

Aug 13

Yuki Shizuya
in
Intuition

Statistics: Multivariate time series analysis — VMA, VAR, VARMA

The mathematical understanding of VMA, VAR, and VARMA and practical Python implementation

Aug 1

Statistics: Multivariate time series analysis — VMA, VAR, VARMA

Aug 1

Yuki Shizuya

Yuki Shizuya

Data Scientist in Japanese IT company. I write blogs about machine learning/deep learning/statistics. https://www.linkedin.com/in/yuki-shizuya-8b4108287/

Following

Help
Status
About
Careers
Press
Blog
Privacy
Terms
Text to speech
Teams