Training / 日本テクノセンター / Japan Techno Center
異常検知AI技術の基礎と応用の実践ポイント:異常の判断を説明可能とする波形可視化技術 <オンラインセミナー>
Define classification loss/errors and limits of normal/abnormal labels · Model normality with neighborhoods,density and one-class detection · Use autoencoder latent representations and reconstruction · Evaluate counterfactual validity,minimal change and plausibility · Apply shapelet learning to scarce-anomaly industrial time series ·…
Visit the provider ↗The details that shape the choice
- Audience
- Ask who the session is designed for
- Format
- Ask about delivery format
- Duration / workload
- 16Nov2026 10:00–17:00 schedule,break/livehours unknown
- Group size
- Ask about participant limits
- Delivery
- Provider-advertised facilitated course route; exact cases, lectures, exercises and live-mode claims remain separately scoped. Two named lecturers explain mathematical foundations,techniques and industrial examples; independently completed waveform model unknown.
- Service area / admission
- Confirm eligibility for the selected learner or group
- Tools
- Ask which tools are used
- Price / charging basis
- Confirm the current price and charging unit with the provider
- Spoken teaching language
- Confirm the teaching language
- Prerequisites and software
- Requirements: High-school mathematics
- ML basics helpful
More from 日本テクノセンター / Japan Techno Center. Confirm current dates, availability and the full fee with the provider.
What the training covers
- Define classification loss/errors and limits of normal/abnormal labels
- Model normality with neighborhoods,density and one-class detection
- Use autoencoder latent representations and reconstruction
- Evaluate counterfactual validity,minimal change and plausibility
- Apply shapelet learning to scarce-anomaly industrial time series
- Relate local attribution and counterfactual waveform explanations
- Generate normalizing counterfactual waveforms and interpret equipment-diagnostic examples
Ask for the full agenda and the time allocated to each activity.
What learners practise
Ask the provider what participants do themselves, what is demonstrated and how feedback is given.
These are descriptions of advertised instruction. They do not establish that individual learners completed the work or achieved a particular result.
Before you book
Ask about the actual session.
Confirm the date, teaching language, participant limit, account requirements and the total fee for the selected route. Ask which exercises learners perform and how feedback is given.
Conditions to check
High-school mathematics required; prior ML helpful. Constitutive counterfactual/time-series analysis retained; no operational reliability or explanation validity certification. Provider-advertised instruction only. Independent completion, assessment, quality, scientific/clinical/legal/financial/safety validity, software/API costs/entitlements, employer input rights and artifact IP unverified.Advertised follow-up
Linked dated policy says postal paper materials/no electronic copy, attendee recording/screenshots prohibited; provider replay and current platform/material logistics unverified.Provider information
A source date describes when information was observed; it is not a new availability check.
Explore this training’s topics and formats
Questions for your shortlist
Related options to investigate
No closely tagged alternatives are recorded for this entry. Browse all training to compare other options.