A Mechanism-Guided Framework for Prioritizing Membrane-Interaction Anti-Vibrio Peptides from Peptidomics Data
A mechanism-informed framework for prioritizing anti-Vibrio peptides from complex peptidomics data.
WARIN RESEARCH LAB · BANGKOK
Ideas without disciplinary borders.
From robotics and evolutionary computation to bioinformatics, intelligent platforms, games, and education - we transform curiosity into systems that can be tested, understood, and used.
My research has never remained inside a single disciplinary boundary. It began with robotics and spatial reasoning, evolved through evolutionary computation and combinatorial optimization, and later expanded into bioinformatics, artificial intelligence, recommender systems, games, education, and human-centred technology.
Across these domains, one question remains constant: How can we transform complex data, uncertain environments, and human needs into systems that are intelligent, understandable, and genuinely useful?
Many projects grow through collaboration with students. An idea may begin in a classroom, become a working prototype, develop into a careful experiment, and eventually contribute to a publication or a longer research programme.
SELECTED WORK · 2025
Research you can explore, not just cite.
Six interactive, first-party research showcases with the problem, method, evidence, and implications in one place.
01
Protein language models
Neuropeptide Classification at Scale
Open showcase ↗ 02 Geospatial AIReflectanceGAN
Open showcase ↗ 03 Computational geometryRobust Voronoi Partitioning
Open showcase ↗ 04 Responsible computer visionUser-Centric Video Privacy
Open showcase ↗ 05 Human-centred AIPractical Open-Set Face Recognition
Open showcase ↗ 06 Thai NLP & transformersGoogle Play Review Quality
Open showcase ↗RESEARCH DOSSIERS · 2024
The research year, opened up.
Seven bilingual dossiers trace each project from its motivating question through method, evidence, findings, and limitations.
Open the complete 2024 collection ↗
01
Computational geometry
Approximate Axis-Aligned Nearest Neighbor
Read dossier ↗ 02 Thai NLP & visual analyticsFiction Gene
Read dossier ↗ 03 Evolutionary game AIEvolutionary Tower Defense Waves
Read dossier ↗ 04 Procedural content generationDeferent: Rogue-Lite Tower Defense
Read dossier ↗ 05 Adaptive game designRopburi Evolutionary PCG
Read dossier ↗ 06 Computer vision & media trustHuman vs. AI Image Authenticity
Read dossier ↗ 07 Thai NLP for educationThai Admissions Sentiment
Read dossier ↗RESEARCH DOSSIERS · CRETECH · ELF ASIA · DOCTORAL
Nine studies, opened from question to evidence.
Explore bilingual, citation-ready dossiers covering intelligent learning, recommender systems, digital platforms, financial machine learning, and evolutionary computation. Each dossier includes method anatomy, evidence boundaries, publication metadata, and concrete routes for student research.
Open the complete dossier collection ↗
06
CreTech 2024
Applied AI and intelligent web systems
Explore six studies ↗ 02 eLearning Forum Asia 2025Inclusive AI and music learning
Explore learning research ↗ 01 Doctoral researchNegative learning in combinatorial optimization
Open doctoral dossier ↗ 06 Supatcha research collectionRNA, protein and peptide intelligence
Explore computational biology dossiers ↗Bioinformatics & Computational Biology
We combine machine learning, biological features, ensemble models, and protein language models to discover meaningful patterns in RNA, proteins, and bioactive peptides - with an increasing emphasis on biological interpretability.
EnsembleNPPred: A robust approach to neuropeptide prediction and recognition using ensemble machine learning and deep learning methods
An ensemble of classical and deep-learning models for robust neuropeptide recognition.
Biologically Interpretable Machine Learning Framework for Neuropeptide Classification
A framework that connects neuropeptide predictions to interpretable biological evidence.
Neuropeptide Classification at Scale with Protein Language Models: Hard Negatives and Cluster-Aware Splits
Protein language models evaluated through hard negatives and cluster-aware data splits.
Ensemble-AHTPpred: a robust ensemble machine learning model integrated with a new composite feature for identifying antihypertensive peptides
An ensemble model with composite features for identifying antihypertensive peptides.
Ensemble of multiple classifiers for multilabel classification of plant protein subcellular localization
A multilabel ensemble for predicting where plant proteins function inside cells.
Weighted Ensemble for Plant Protein Subcellular Localization Using Particle Swarm Optimization
Particle swarm optimization learns classifier-localization weights, with a live local-inference lab exposing evidence before weighting.
Human Protein Subcellular Localization · Local PSO Lab
A human-specific experiment combining Human Protein Atlas evidence, local predictors and transparent multilabel PSO voting.
mSRFR: a machine learning model using microalgal signature features for ncRNA classification
A machine-learning model using microalgal signature features for ncRNA classification.
Application of Random Forest in Limited Size Human Long Non-coding RNAs Identification with Secondary Structure Features
Random forests and RNA structural features for useful learning from limited biological data.
AdaBoost Algorithm with Random Forests for Plant and Animal Precursor MicroRNAs Classification
A boosted random-forest approach for plant and animal precursor microRNA classification.
Identification of Plant Precursor miRNAs using Structural Robustness and Secondary Structures Features
Structural robustness reveals precursor-miRNA signals beyond sequence information alone.
Enhanced Viral Precursor MicroRNA Identification with Structural Robustness Features in Back-Propagation Neural Network
Structural-robustness features strengthen neural identification of viral precursor microRNAs.
Viral MicroRNA Precursors Discovery with Self-containment Index and Derivative Features using Back-propagation Neural Network
An IJSST 2016 study using self-containment features to identify viral miRNA precursors.
RNA secondary structure prediction with coincidence algorithm
Combinatorial optimization applied to the complex search space of RNA secondary structures.
Improving the Performance of BLAST in a Memory Limited Environment
Making biological sequence search more practical under restricted memory.
Evolutionary Computation & Optimization
This foundational direction studies how candidate solutions can learn, preserve diversity, and navigate large search spaces with multiple objectives and constraints.
Multi-objective combinatorial optimisation with coincidence algorithm
A foundational extension of the Coincidence Algorithm to multi-objective combinatorial problems.
Coincidence Algorithm for Combinatorial Optimisation and Its Applications
An overview of the Coincidence Algorithm and its combinatorial applications.
Negative Correlation Learning in the Estimation of Distribution Algorithms for Combinatorial Optimization
Negative correlation encourages complementary probabilistic search behaviours.
The effectiveness of hybrid negative correlation learning in evolutionary algorithm for combinatorial optimization problems
An evaluation of hybrid correlation learning in evolutionary optimization.
Hybrid positive and negative correlation learning in estimation of distribution algorithm for combinatorial optimization problems
Balancing cooperation and diversity among probabilistic models.
Solving multimodal combinatorial puzzles with edge-based estimation of distribution algorithm
An edge-based EDA designed to discover several high-quality puzzle solutions.
Solving multimodal problems by coincidence algorithm
Maintaining multiple promising regions in multimodal optimization.
Solving Sudoku puzzles with node based Coincidence algorithm
Node-based combinatorial reasoning demonstrated through Sudoku.
Application of Node Based Coincidence algorithm for flow shop scheduling problems
Node-based search applied to flow-shop scheduling.
Application of estimation of distribution algorithms for solving order acceptance with weighted tardiness problems
Balancing order selection, weighted tardiness, and production constraints.
The merging of MPS and order acceptance in a semi-order-driven industry: A case study of the parasol industry
Connecting master production scheduling and order acceptance in a real industrial case.
Overtime capacity expansion in order acceptance with node based estimation of distribution algorithms
Deciding when overtime capacity creates enough value to accept more orders.
Application of node based estimation of distribution algorithms for solving order acceptance with capacity balancing problems by trading off between over capacity and under capacity
Balancing the risks of excessive and insufficient production capacity.
Application of node based coincidence algorithm for solving order acceptance with multi-process capacity balancing problems
Coordinating order acceptance across interconnected production stages.
An Optimization Model and Algorithms for Order Acceptance Problem with Overtime Strategy
A formal optimization model for order acceptance with overtime strategy.
Gaussian quantum-behaved particle swarm with learning automata-adaptive attractor and local search
A swarm optimizer combining adaptive attractors, Gaussian movement, and local refinement.
A nondominated adversarial search algorithm for a three-player chess game
Nondominated adversarial search for strategic decisions in three-player chess.
Approximate Axis-Aligned Nearest Neighbor
An efficient geometric method for approximate nearest-neighbour retrieval.
Robust Voronoi Partitioning for Scaffold Architecture via Learned-σ Centroids and Stable Normal Flow
Learned centroids and stable normal flow for robust scaffold partitioning.
Applied AI, Vision & Intelligent Systems
We turn machine learning into practical systems for authenticity, privacy, engineering, geospatial analytics, healthcare, and decision support.
ReflectanceGAN: Geospatial SAR-to-MSI Translation for Cloud-Agnostic Sentinel-2 Analytics
Translating radar imagery into multispectral representations when clouds obstruct optical observation.
Automated Bounding Dimension Retrieval: A Machine Learning-Driven Framework for Engineering Drawing Interpretation
Machine learning for extracting dimensional information from engineering drawings.
Human vs. AI: Leveraging Machine Learning and Deep Learning to Verify Image Authenticity
Distinguishing authentic visual content from AI-generated imagery.
Real-time Car Part Instance Segmentation: the Comparison of the State-of-the-Art
Comparing modern instance-segmentation methods for real-time car-part recognition.
User-Centric Video Privacy Preservation: Automated Face Detection, Recognition, and Blurring for PDPA and Beyond
A user-centred video privacy pipeline designed for PDPA-aware workflows.
Practical Open-Set Face Recognition: Operator-Centric Design with KD-Tree Search
Open-set face recognition designed for real operators and unknown identities.
A Real-Time Pressure Ulcer Prevention Monitoring System for Elderly Care Using Activity-Based Detection
Activity recognition transformed into preventive support for elderly care.
Recommender Systems & Human-Centred Platforms
These projects combine language models, collaborative filtering, similarity search, UX, and usability testing to help people make better everyday decisions.
Text-based Movie Recommendation Web Application with DistilBERT
DistilBERT, cosine similarity, and KNN make 45,000 movies searchable through natural-language preferences.
Development of a Web-Based Anime Streaming and Recommending Application
A Naive-Bayes anime recommender combined with real-time community interaction.
Hybrid Property Post Recommendation System
A hybrid KNN-SVD property recommender addressing cold starts and sparse data.
B-Trade Innovating Book Exchange Through Web Technology
A sustainable book-exchange community enhanced by personalized recommendations.
MaidConnect - Innovative Connection Between Housemaids and Employers
KD-tree matching connects housemaids and employers through proximity and trust.
Automated Trading Bot Using Hidden Markov Models
Hidden Markov Models identify latent Bitcoin market states for automated trading decisions.
Fiction Gene: A Thai Genre Based Fiction Visualization Testimonial System
A visual system for exploring the genre characteristics of Thai fiction.
Thai MBTI Psychographic Segmentation: Complement Naïve Bayes vs. Thai Transformer on Short Social Media Posts
Classical and transformer models compared for personality-oriented Thai text segmentation.
Google Play Review Quality Scoring for Digital Engagement and App Development Using Transformer Models
Transformer models score review quality rather than treating every comment equally.
Enhanced Sentiment Detection in Thai University Admissions Using Complement Naïve Bayes
Thai sentiment analysis focused on university-admission experiences.
Analysis of Deep Demographics for Advertisement Recommendation
Richer demographic representations for more relevant advertisement recommendations.
Games, Creativity & Procedural Intelligence
Games become a laboratory for creativity, adaptation, and search, with evolutionary computation acting as a design partner.
Procedural Content Generation for 2.5 D Rogue-Lite Games: An Evolutionary Algorithm Approach
Evolutionary generation of replayable 2.5D rogue-lite content.
Automating Tower Defense Game Level Design with Evolutionary Algorithms
Searching for tower-defense levels that balance challenge and playability.
Revolutionizing Tower Defense: Automating Game Level Design with Rogue-lite Integration
Automated tower-defense level design enriched with rogue-lite variation.
AI for Education & Human Learning
Educational technology should make difficult ideas visible, create room for experimentation, and help learners and teachers reflect on how learning happens.
VisualRight: A Web-Native Interactive Playground to Democratize AI Literacy through Visualization and Simulation
A browser-based playground for seeing and manipulating AI concepts.
AI-Powered Guitar Learning Platform: Advancing Digital Music Education for Diversity, Equity, and Inclusion
An AI-supported guitar-learning platform designed for accessibility and inclusion.
Robotics & Spatial Intelligence
The research journey began with spatial reasoning and distributed autonomous systems - themes that still echo through today’s intelligent-system work.
Relative Position Localizing System for Multiple Autonomous Mobile Robots in Distributed Robotic System: ---Algorithm to Detect Overlapped SLPM Patterns---
Relative localization among multiple autonomous mobile robots in a distributed system.
WORK WITH ME
Bring a question, a problem, or an unfinished idea.
Students do not need to arrive with a finished research question. Together, we can turn an interest in biology, games, education, language, images, algorithms, or an everyday problem into a researchable question, a working system, and evidence that others can learn from.
Start a conversation