Lately, I've been studying statistical mechanics.

Kernel Logistic Regression Learning for High-Capacity Hopfield Networks Hebbian learning limits Hopfield network storage capacity (pattern-to-neuron ratio around 0.14). We propose Kernel Logistic Regression (KLR) learning. …

My paper has been published. / Kernel Logistic Regression Learning for High-Capacity Hopfield Networks doi.org/10.1587/tran...

My recent work :)

GitHub - tam17aki/music-esprit-python: Python implementations of high-resolution parameter estimation algorithms for sinusoidal signals including MUSIC (Spectral/Root) and ESPRIT. Python implementations of high-resolution parameter estimation algorithms for sinusoidal signals including MUSIC (Spectral/Root) and ESPRIT. - GitHub - tam17aki/music-esprit-python: Python impleme...

My recent work: Python implementations of high-resolution parameter estimation algorithms for sinusoidal signals including MUSIC (Spectral/Root) and ESPRIT.
github.com/tam17aki/mus...

This paper has been accepted to APSIPA ASC 2025.

(PDF) Kernel outlier detection PDF | new anomaly detection method called kernel outlier detection (KOD) is proposed.It is designed to address challenges of outlier detection in... | Find, read and cite all the research you need on ...

My paper "Kernel Random Projection Depth for Outlier Detection" has been cited in this paper. : www.researchgate.net/publication/...

This paper has been accepted by IEICE Trans.

Frequency Estimation Method Using Sinusoidal Constraint Differential Equation (SCDE) and Python Implementation - 備忘録 Introduction Conventional frequency estimation methods like FFT (Fast Fourier Transform) and autocorrelation functions are widely used. However, these methods share a common challenge: it is difficult...

Recently, I posted the article about the frequency estimation with differential equation (a translation from my Japanese article). I think it is interesting that a signal processing problem of frequency estimation reduces to solving an algebraic equation. tam5917.hatenablog.com/entry/2025/0...

Preparing lecture materials about speech recognition and synthesis.

Akira Tamamori: Quantitative Attractor Analysis of High-Capacity Kernel Logistic Regression Hopfield Networks https://arxiv.org/abs/2505.01218 https://arxiv.org/pdf/2505.01218 https://arxiv.org/html/2505.01218

KLR/KRR can be considered as replacing the Hopfield network's recollection rules with those that have high pattern separation ability using kernel methods.

Based on the two kernel method papers that I previously uploaded as preprints, it looks like I will be able to tackle many more research topics.

I've been concerned about the impact of generative AI on the field of education. I teach programming classes, and there are many difficult aspects to it.

Akira Tamamori: Kernel Ridge Regression for Efficient Learning of High-Capacity Hopfield Networks https://arxiv.org/abs/2504.12561 https://arxiv.org/pdf/2504.12561 https://arxiv.org/html/2504.12561

Akira Tamamori: Kernel Logistic Regression Learning for High-Capacity Hopfield Networks https://arxiv.org/abs/2504.07633 https://arxiv.org/pdf/2504.07633 https://arxiv.org/html/2504.07633

Presenting at an international conference abroad is good, but I also think it would be better to stay in the own country and submit and present a journal article. The former is financially expensive and can make you sick.

submitted my paper. waiting for good news.

I was writing a draft of my paper today. I'm glad I managed to fill up the pages.

Energy-Based Models with Applications to Speech and Language Processing Energy-Based Models (EBMs) are an important class of probabilistic models, also known as random fields and undirected graphical models. EBMs are un-normalized and thus radically different from other p...

FYI arxiv.org/abs/2403.10961

I am also interested in "Energy-Based Models with Applications to Speech and Language Processing" www.nowpublishers.com/article/Deta...

Successfully downloaded the recently published monograph / Sound Field Estimation: Theories and Applications www.nowpublishers.com/article/Deta...

I am busy every day preparing lecture materials and reviewing papers.

Next, I'm studying statistical mechanics related to Ising model.

I'm studying logistic regression.

『Mel-Spectrogram Inversion via Alternating Direction Method of Multipliers』に基づくメルスペクトログラムからの音声復元手法をPythonで実装してみた - 備忘録 はじめに 最近，以下の論文がarXivで見つかった： Yoshiki Masuyama, Natsuki Ueno, and Nobutaka Ono, "Mel-Spectrogram Inversion via Alternating Direction Method of Multipliers, " arXiv:2501.05557, 2025 arxiv.org メルスペクトログラムから...

This is the article in my blog (in Japanese). I hope you can translate it. tam5917.hatenablog.com/entry/2025/0...

iPALM-based mel-spectrogram inversionを活用してMFCCからの音声復元を試してみた - 備忘録 はじめに 前回記事の続きである． tam5917.hatenablog.com ところでlibrosa には mfcc_to_audio という関数が用意されている． librosa.org この関数はMFCC（音声認識では定番の特徴量）を音声に復元する機能を実装している．処理の具体的な中身は mfcc_to_mel (逆DCT→dBスケールからメルスケールへの変換) mel_to_audio と...

This is the article in my blog (in Japanese). I hope you can translate it. tam5917.hatenablog.com/entry/2025/0...

Emacs is my favorite editor.

I have updated my Emacs to the latest ver30.1. www.gnu.org/software/ema...

Also I found his doctoral advisor was John Hopfield. mathgenealogy.org/id.php?id=12...

This book is truly a masterpiece.