Summary of "WEBINAR NASIONAL: Masa Depan Computer Science: AI, Data, dan Otomatisasi di Era Society 5.0"
Main ideas and lessons conveyed (by topic)
1) Opening context and overall theme (Society 5.0 + CS)
- The webinar’s theme is the future of computer science, centered on AI, data, and automation in the Society 5.0 era.
- Society 5.0 is presented as a shift toward human-centered digital transformation by integrating cyberspace and physical space.
- Key “pillars” repeatedly referenced:
- AI (intelligent technologies; AI literacy is increasingly important)
- Data science / big data (data as the foundation for prediction and decision-making)
- Automation & robots (more capable systems that still must be governed by humans)
2) Computer science is broader than programming
A central message: computer science is not only coding. It is an ecosystem including:
- data pipelines,
- analytics and modeling,
- governance and ethics,
- cybersecurity,
- and human decision-making.
Methodologies / structured concepts mentioned (detailed bullets)
A) Biomedical Machine Learning for Health Transformation (Prof. Imam Tahyudin)
Context
- Society 5.0 places humans at the center of digital transformation.
- Biomedical ML is framed as AI that learns from data to enable predictive/classification tasks without requiring explicit programming.
ML types referenced
- Supervised learning
- Unsupervised learning
- Reinforcement learning
- Deep learning (described as a fast-developing branch)
Biomedical ML challenges
- High heterogeneity of medical modalities (record types like imaging, ECG, EEG have different characteristics)
- Curse of dimensionality (features ≫ samples) → overfitting risk
- Class imbalance (critical conditions like death/septic shock are rarer than normal cases)
- Missing data (records are often incomplete)
- Time-series non-stationarity/nonlinearity (ECG/EEG dynamics)
Research/solution directions described
- Working with limited medical data
- Example approach: data augmentation + transfer learning for small-scale datasets
- Biomedical deep learning for practical monitoring
- Example: temperature detection using bioelectric/potential-based data
- Stroke-related predictive systems
- Build an app to predict stroke risk within ~10 years using multiple variables (e.g., hypertension, diabetes, heart disease, smoking, high cholesterol, etc.)
- Use an expert system approach to capture neurologists’ decision/therapy logic
- Mortality and hospitalization length prediction
- Aim: support clinical teams and align incentives/coverage constraints (discussion references BPJS “coverage days” logic)
- Imaging classification
- Classify stroke types from brain images (ischemic vs hemorrhagic vs transient ischemic attack)
- IoT-based early detection
- Early stroke prediction using physiological signals (heart rate, oxygen saturation, body temperature)
- Mentions GPS-based fast handling and the “golden time” concept
Key obstacles for AI in Indonesian healthcare
- Data infrastructure gaps
- Regulatory/standards issues for AI diagnostic tools (e.g., unclear certification alignment with medical record systems)
- Human resource gaps (IT and hospital staffing)
- Privacy and data sovereignty concerns
B) “From Data to Paper” using Big Data & Data Science for Publications (Prof. Suyoto)
Definitions and building blocks
- Big data: very large, complex, diverse datasets typically described by 5Vs
- Volume (very large data)
- Velocity (real-time generation/processing)
- Variety (structured, unstructured, semi-structured)
- Veracity (data accuracy/reliability)
- (A value-related “V” may also appear, though the transcript emphasizes the above)
- Data science: cross-disciplinary process to extract insights from data (statistics + computation + machine learning)
- Dataset: a structured collection of data used as input for analysis/modeling
Core workflow for making a publishable paper
- Obtain raw data (streaming/web/social media/sensors/IoT/biomedical sources)
- Process into clean, structured datasets
- Analyze/model with ML/AI to generate findings and visualizations
- Use results to support predictions/decision-making
- Publication emphasis:
- Clean dataset → stronger analysis → easier publication
- Scopus trend logic is referenced (e.g., citation/topic relevance)
Practical publishing strategy (tools & approaches mentioned)
- Use Scopus to search topic trends (keywords + filtering strategies)
- Use mapping/visualization tools such as Vosviewer to identify research gaps and opportunities
- Consider publication venues:
- Sinta journals (national accreditation tiers)
- Scopus-indexed international seminars/proceedings
- Mentions “free publication” platforms such as noapc.com (with caution about continuity/coverage)
Audience-relevant Q&A guidance (examples)
- Finding open/free datasets for students:
- Start with available datasets on Kaggle
- Ensure the data suits the research goals (not only “coding practice”)
- Data augmentation example:
- Rotating and flipping image angles (e.g., leaf images) to create additional training examples
C) Decision Support Systems (DSS) in Society 5.0: Integrating AI, Data, Automation (Prof. Sri Andayani)
DSS definition and purpose
- A DSS is a computer application that supports policy formulation and structured decision-making.
Human-centric shift
- AI should not replace humans; it should function as a co-worker/assistant in decision-making.
DSS characteristics described
- Uses analytical algorithms and computing technology
- Combines:
- historical data,
- analytical models,
- and user preferences
Example
- GPS-like decision assistance:
- suggests fastest routes,
- but the human still steers and decides.
Multi-Criteria Decision Making (MCDM)
- MCDM: select the best alternative among multiple options with conflicting criteria
- Weigh criteria and compare alternatives
- Multiple stakeholders may create multiple decision makers
Data-Driven Decision Making (DDDM)
- DSS evolves from static data to real-time data-driven decision making
- Emphasized principles:
- data-driven + evidence-based decisions
- continuous/iterative and transparent integration of data
Explainable AI (XAI)
- Challenge: black-box recommendations without reasoning reduce trust
- XAI provides logical reasons to improve user confidence
Reported performance benefits (as stated)
- Decision accuracy improvement (~15–20% cited)
- Processing time reduction (~20–30% cited)
Challenges to real-world adoption
- Data quality issues (messy/unstructured)
- Algorithmic bias and misalignment with human values
- Skills crisis (need for HR/people literacy)
- Trust remains the biggest barrier
D) Bioinformatics as a New Beginning for Computer Science Research (Prof. Wisnu Ananta Kusuma)
Core idea
- Bioinformatics integrates computer science, mathematics, statistics, and biology to accelerate research in health, agriculture, environment, and pharmaceuticals.
Computational thinking emphasized
- Abstraction
- Decomposition
- Pattern recognition
- Algorithmic problem solving (sequential algorithms)
How biology inspires algorithms
- Genetic algorithms inspired by evolution for optimization
- Ant colony / swarm-like methods inspired by ant search behavior
DNA data representation
- Encode DNA bases as characters (A, T, G, C)
- Large data volumes lead to “data explosion,” requiring computation
Sequence comparison / dynamic programming
- Mentions common subsequence and dynamic programming as analogies for comparing genetic sequences
- Use cases:
- comparing virus similarity
- speeding drug repurposing ideas
Bioinformatics research workflow described (drug discovery emphasis)
- Detect associations between diseases
- Identify targets (proteins involved)
- Functional enrichment to understand roles in metabolism pathways
- Predict compound–target interactions
- Use molecular docking/simulation to evaluate binding
Algorithms/tools mentioned
- Community detection approach (pseudo-village / modularity style)
- Community centrality variant (new evaluation matrix)
- Predictive models including:
- SVM
- multilayer perceptron (deep learning component)
- random forest / ML ensemble
- Transfer learning for peptide–cancer target prediction
- Modeling herbal medicine as multi-entity optimization (problem framing noted as KNPS-like)
- Multi-omic integration and experimental validation (in vitro)
Bioinformatics challenges
- Heterogeneous formats and noise/errors in sequencing
- High dimensionality (thousands to hundreds of thousands of parameters)
- Need for cross-disciplinary collaboration and high computational capability
Closing analogy
- Sequencing/reference data as a blueprint vs individual differences
- National progress depends on combining individual competencies (synergy theme)
E) Optimization using “Metaheuristics” in Engineering Field (Dr. Purbadaru Kusuma)
Metaheuristics purpose
- Used for optimization when exact solutions are too expensive or impractical.
- Applied across:
- engineering,
- ICT/AI searching,
- scheduling,
- routing,
- allocation,
- finance,
- and more.
Examples of application domains described
- Power systems
- economic emission dispatch (optimize cost + emissions)
- Agriculture / precision farming
- crab cultivation optimization in vertical shelving settings
- Mechanical design
- welded beam, spring design
- Production and allocation
- order allocation (supplier distribution)
- Finance
- portfolio allocation for IDX30 blue-chip stocks
- Outsourcing decisions
- outsource choice + quantity planning
- COVID-era routing and scheduling
- coordinated ambulance routing
- pickup/delivery routing for collaborative city couriers (multi-operator coordination)
- Vaccination scheduling
- capacity-constrained scheduling for vaccination centers
- Education MBKM scheduling model
- assignment/course matching model
- Production scheduling
- flow-shop scheduling with multiple machines per stage
- Operations research connection
- derived from classic problems like TSP, vehicle routing, pickup-delivery, and clustering-like variants
Practical note
- Modern AI tools may make coding easier, but understanding the model and the problem remains crucial.
F) Future of Computers: AI, Data, Automation in Society 5.0 (Prof. Hari Sutanto)
Paradigm shift
- Society 5.0 emphasizes human-centered quality of life
- Contrast: Society 4.0 emphasized industrial automation/productivity
Integrated “brain–data–action” view
- Data → AI processing/learning → automation/action execution
- Data and automation are treated as a connected discipline, not separate fields
Technological enablers
- AI, big data, IoT, robotics, cloud computing, cybersecurity
- Mentions multimodal AI (combining text/image/audio)
Human role and trust
- AI should be a partner, not a replacement
- Concern about AI “taking over” is framed as depending on how humans choose to use it
AI workflow cycle described
- Data → information/patterns → knowledge → machine learning → automatic actions → model updates with new data
- Progression mentioned:
- manual methods → rule-based automation → intelligent automation (AI-driven)
Explainability and governance
- Maintain critical thinking; avoid “blind trust” due to speed
- “Garbage in, garbage out” emphasized:
- wrong inputs lead to wrong outcomes
- Ethics/legal/privacy concerns
- For healthcare, the tolerance for error is described as near-zero
Q&A concept
- Overfitting mitigation ideas:
- add data diversity,
- preprocessing and augmentation,
- ensure the model doesn’t learn only uniform patterns
G) Health 4.0 Digital Revolution in Health Services (Dani Sasmoko)
Health evolution
- Health 1.0: manual records on paper
- Health 2.0: data entry into spreadsheets
- Health 3.0: integrated/online history (e.g., BPJS-based access)
- Health 4.0: AI + IoT + big data + predictive/proactive monitoring
- Health 5.0: holistic integration across physical/digital/biological health ecosystem
AI use cases in hospitals
- Personalized risk identification (no universal “100% algorithm”; healthcare vulnerability varies)
- Proactive/predictive services using integrated data and remote access
- IoT-based medical devices and “smart hospital” direction (smart beds, smart environments)
- Telemedicine/telesurgery (remote consultation and procedures)
- Example application categories:
- wound/diabetic severity prediction using image shape
- cancer stage prediction using AI
- mortality risk prediction and hospitalization needs
- real-time health monitoring signals (heart rate, oxygen/breathing patterns, etc.)
- Mentions blockchain in healthcare and platform/country references (e.g., HelloDoc)
Challenges in Indonesia
- Cybersecurity threats and attacks
- Uneven internet availability across islands
- Unclear or insufficient regulations
- Ethical violations
- Low digital literacy for health applications
Proposed strategy
- A government + private sector roadmap
- Stronger human resource development
- Multi-party collaboration
H) Closing: Society 5.0 and Future World of IT, AI, Data, Automation (Dr. Doni Novelendri)
Society 5.0 definition
- Introduced by Japan (2016) to address issues not solvable by Industrial 4.0:
- aging population,
- economic hardship,
- climate change
- Key feature: deep integration of physical + digital worlds
Role of AI
- Optimizes processes previously requiring human effort
- Supports intelligent decision-making via ML from historical data
- Example: AI assistance for doctors/lab diagnosis (dissertation on liver cancer detection)
Role of data
- Data is described as equivalent to oil/gas in value
- Competitive advantage comes from collecting/storing/managing data well, then processing it via data mining/statistics
IT integration and automation loop
- Data collected from physical systems (e.g., traffic sensors) enables real-time analysis and actions
Opportunities and job roles listed
- Data scientist
- Machine learning/AI engineer
- IoT + systems developer
- Cybersecurity specialist
Challenges/risks listed
- Ethics and bias in algorithms
- Accountability (e.g., autonomous vehicle accident responsibility)
- Privacy issues (large personal data needs, identity-linked systems)
- Data security/data leakage risks
- Digital divide (unequal connectivity/devices across regions)
Speakers / sources featured (identified people)
- Paulina Kinanti Eka Praning Tias — MC/moderator (webinar guide)
- Dr. Josep Teguh Santoso, MKom — Chancellor of Tekom University Semarang (opening remarks)
- Mr. Budi Hartono, SKOM, MKOM — Head of Undergraduate Program, Informatics Engineering at Tekom University Semarang (welcoming speech)
- Prof. Dr. Eng. Imam Tahyudin, S.Si., MKOM, M.M. — Amikom University Purwokerto (Biomedical Machine Learning)
- Prof. Ir. Suyoto, M.Sc., Ph.D. — Atma Jaya University Yogyakarta (Data to Paper / big data & data science)
- Prof. Dr. Sri Andayani, S.Si., M.Kom. — Yogyakarta State University (Decision Support Systems)
- Prof. Dr. Eng. Wisnu Ananta Kusuma, ST., MT. — IPB University (Bioinformatics)
- Dr. Purbadaru Kusuma, ST., MT. — Telkom University Bandung (Metaheuristics optimization)
- Prof. Ir. Hari Sutanto, SKOM, M.Sc. — Budi Luhur University Jakarta (Future of Computers in Society 5.0)
- Dani Sasmoko, ST, M.Eng. — Tekom University Semarang (Health 4.0 digital revolution in health services)
- Dr. Doni Novelendri, SKOM, MKOM — Padang State University (Closing: future world of IT/AI/data automation)
Participant questions (identified)
- Januar Tito — participant question (asked during Prof. Suyoto and later in Prof. Wisnu’s Q&A)
- Muhammad Firman Gani — participant question (overfitting question during Prof. Hari’s Q&A)
- Januar Tito (again as “Brother Januar Tito”) — participant question (DNA sequence assembly comparison during Prof. Wisnu’s Q&A)
- Januar “Mas Yanar/Yanuar” — participant question about open/free datasets for final assignment (answered by Prof. Suyoto)
Institutions/organizations mentioned as organizers/collaborators include: Tekom University Semarang, Amikom Purwokerto, Atma Jaya Yogyakarta, Yogyakarta State University, IPB University, Telkom University, Budi Luhur University, Padang State University, STSTKOM, PerkIFI, PTIC, toploker.com, and others referenced in the transcript.
Category
Educational
Share this summary
Is the summary off?
If you think the summary is inaccurate, you can reprocess it with the latest model.
Preparing reprocess...