[Web Seminar] Tips for Annotation Management to Successfully Implement AI Machine Learning - Based on Experience Creating Over 48 Million Data Annually - (10/8)

September 11, 2020

[Web Seminar] Tips for Annotation Management to Successfully Implement AI Machine Learning - Based on Experience Creating Over 48 Million Data Annually - (10/8)

AI Machine Learning Annotation Seminar on October 8, 2020 (Thursday)

 

[Web Seminar] Tips for Successful Annotation Management in AI Machine Learning

~ Over 48 million data creation experience for years ~


will be held.


———————————————————————————————————————


In AI/machine learning, the quality of annotated training data significantly affects the accuracy of AI.

To improve accuracy, it is necessary to have a large number of human resources for creating a vast amount of training data, making quality and talent management extremely important.

 

    While tools and their usage are important for annotation, there are few areas that can be automated, and manual work cannot be avoided.

    Therefore, the weight that depends on the characteristics and abilities of the annotators is significant, and no matter how much manualization or standardization is implemented, if the management points are misaligned, productivity and quality will not improve.

    Therefore, the improvement of human performance and the "appropriate balance of coexistence" of manuals greatly influence the quality and productivity of annotations, but this is surprisingly overlooked.


    Human Science creates over 48 million training data sets annually.

    In this seminar, we will focus on the quality management aspect of annotations based on that knowledge, highlighting the important points.


  • 1. How to incorporate information into the manual
  • ②Mechanisms and devices that maximize human performance

    • We will introduce examples and case studies.


    This is a seminar that we would like those in charge who have the following challenges to participate in!


  • ・I created manuals and guidelines, but I'm not sure how much information I should include (if I include too much, it will never end).
  • ・Although I created the manual, the questions and interactions from the annotators do not decrease.
  • ・I am told by the annotator that balancing quality and productivity is impossible, and I am troubled by this.
  • ・Unable to manage a large number of annotators
  • The productivity varies greatly depending on the annotator, and we are struggling as there has been no improvement despite guidance.
  • ・I want to focus on development work, so I can't manage the annotators (annotation).

    • After the seminar, we will provide individual consultations as needed.

    Please check the "Request for Individual Consultation" box on the application form.

     

    Feel free to join us!

Program

  • – Main Seminar (50 minutes)
  • – Q&A (10 minutes)
  • – Individual Consultation Session (First 2 Applicants)

Eligible Persons

  • ・Those who are working on AI implementation, PoC, and other annotations in their own projects
  • ・Those considering BPO for annotation
  • - Those who feel challenges with their company's annotation process
  • ・Those who are managing annotations

 

  • *If the number of applications exceeds the capacity, a lottery may be held.
  • *We allow up to two participants from the same company.
  • * If there are many participants, a lottery will be held.

Event Guidelines

Date and Time
  • October 8, 2020 (Thursday) 16:00 - 17:00
  • Individual Consultation (for those who wish to participate): 17:00 - (Conducted via web conference)
Cost
Free
How to Participate
  • This seminar is a web seminar using Zoom.
    Details on how to participate will be communicated separately via email to those who have registered.
  • For information about the playback environment for Zoom, please check here.

Application

  • Please apply for participation in this seminar using the form below.
  • We will get back to you from our representative.

 

Application Deadline: Monday, October 5, 2020

 

  • Participation from competing companies and individuals may be declined.
  • * The event may be canceled if the minimum number of participants is not reached.
  • We will contact you about one week in advance in case of cancellation.
  • We will contact participants via email a few days before the event regarding the method to view the webinar and the access URL.
  • Individual consultations will be conducted in a web conference format.

Instructor

IMG_3577_3

Yasuhiro Suzuki

  • ・Annotation Project Manager
  • As a leader of a global IT help desk in my previous job, I was involved in management tasks such as training domestic and international members and performance management, as well as document creation and improvement for overseas data center transfers and translation.
  • ・In my current position, I leverage my experience and knowledge in translation and documentation work, having worked as a localization engineer and consultant focused on machine translation. Currently, I am responsible for project management of large-scale natural language annotation projects (including English-Japanese localization) on-site at client locations.

Manabu Kitada

  • ・Annotation Project Manager
  • Since the establishment of the Annotation Department, we have gained extensive experience in various annotation tasks and management, focusing on natural language processing, including client-resident POC projects, remote projects, and large-scale onsite projects.
  • - In project management and team leadership, we have planned and implemented numerous initiatives aimed at enhancing communication among annotators and promoting knowledge sharing.
  • ・Currently responsible for managing image annotation related to the automotive industry.

Application

  • Please apply for participation in this seminar using the form below.
  • We will get back to you from our representative.

 

Application Deadline: Monday, October 5, 2020

 

  • Participation from competing companies and individuals may be declined.
  • * The event may be canceled if the minimum number of participants is not reached.
  • We will contact you about one week in advance in case of cancellation.
  • We will contact participants via email a few days before the event regarding the method to view the webinar and the access URL.
  • Individual consultations will be conducted in a web conference format.

Scheduled Seminars


Details will be posted later. If you would like to receive a seminar notification email, please register using the button below.



Contact Us

If you have any questions about the content of this seminar, please contact hsweb_inquiry@science.co.jp .

For inquiries by phone, please contact us at the number below.

TEL: +81 35-321-3111

Person in Charge: Business Promotion Department, Saga