Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.
1
Auto Content Moderation
in C2C e-Commerce
Shunya Ueta, Suganprabu Nagarajan, Mizuki Sango
(Mercari, inc)
2020 USENIX Con...
2
1. Content Moderation
2. Auto Content Moderation in C2C e-Commerce
3. Task design and model strategy
4. Offline/online ev...
3
Identify potentially unsafe or inappropriate content in service
● App Discovery with Google Play, Part 3: Machine Learni...
4
The Mercari app is a C2C marketplace where individuals can easily sell used items
What is Mercari?
Japan
U.S.
Monthly ac...
5
Why Content Moderation in C2C e-Commerce?
C2C e-Commerce
Sellers Buyers
We want to decrease risk for customer and market...
6
Content Moderation system
C2C e-Commerce
Sell items Discover
Moderator
Manual review
Moderation
Service
Hide items & Ale...
7
Concept of Moderation Service: Rule based
Moderation
Service
Rule based
Pros
● Easy to develop and can be
quickly releas...
8
Concept of Moderation Service: ML
Moderation
Service
Rule based
Pros
● Automatically learns the features
of items delete...
9
How to create the data for ML
Rule based
Moderator
Machine
Learning
Sell items
Report items
Hide items & Alert
Positive
...
10
Task Design
● Data is highly imbalanced
● Each violated topic’s total
number of alerts is bounded
by moderator team
All...
11
Multimodality of content
Case of items
Items have multimodal data
● Image
● Text
● Category
● Brand
● Price, etc.
We us...
12
Model selection based on dataset size
● Gradient Boosted Decision Trees (GBDT)
→ Efficient for training and inference wh...
13
Offline evaluation
Metric is Precision@K: K is the bound on the daily total number of alerts
in each violated topic deci...
14
Online evaluation
→ Faster decision making leads to efficient operation
Current
Model
New
Model
Same traffic
Moderator
Ma...
15
Offline/online evaluation result
Algorithms Offline Online
GBDT +18.2% Not Released
GMU +21.2% +23.2%
Table shows the re...
16
Container based Training Pipeline
Data Load
Write manifest files containing requirements like CPU, GPU and Storage
CPU C...
17
Serving system architecture
Message
queue
Message
queue
proxy layer prediction layer
.
Preprocessing + inference
Contai...
18
Horizontal Pod Autoscaler by kubernetes
● Reliable system: Traffic changes with time,
HPA can adopt to varying traffic
● ...
19
Impact of Machine Learning system
Discovered 100 violating items
Moderator
Manual review
Moderation
Service
Rule based
...
20
If you have a question to this talk
First author is Shunya UETA, please e-mail: hurutoriya@mercari.com
Acknowledgements...
Upcoming SlideShare
Loading in …5
×

1

Share

Download to read offline

Auto Content Moderation in C2C e-Commerce at OpML20

Download to read offline

2020 USENIX Conference on Operational Machine Learning

Authors:
Shunya Ueta, Suganprabu Nagaraja, and Mizuki Sango, Mercari, inc.

Abstract:
Consumer-to-consumer (C2C) e-Commerce is a large and growing industry with millions of monthly active users. In this paper, we propose auto content moderation for C2C e-Commerce to moderate items using Machine Learning (ML). We will also discuss practical knowledge gained from our auto content moderation system. The system has been deployed to production at Mercari since late 2017 and has significantly reduced the operation cost in detecting items violating our policies. This system has increased coverage by 554.8 % over a rule-based approach.

Related Books

Free with a 30 day trial from Scribd

See all

Related Audiobooks

Free with a 30 day trial from Scribd

See all

Auto Content Moderation in C2C e-Commerce at OpML20

  1. 1. 1 Auto Content Moderation in C2C e-Commerce Shunya Ueta, Suganprabu Nagarajan, Mizuki Sango (Mercari, inc) 2020 USENIX Conference on Operational Machine Learning JULY 27–AUGUST 7, 2020
  2. 2. 2 1. Content Moderation 2. Auto Content Moderation in C2C e-Commerce 3. Task design and model strategy 4. Offline/online evaluation 5. System architecture 6. Business Impact Contents
  3. 3. 3 Identify potentially unsafe or inappropriate content in service ● App Discovery with Google Play, Part 3: Machine Learning to Fight Spam and Abuse at Scale ● YouTube Community Guidelines enforcement ● AI advances to better detect hate speech by Facebook ● Advances in content understanding, self-supervision to protect people by Facebook ● Facebook Transparency Report ● A Safe and Secure Marketplace by Mercari ● etc. Content Moderation
  4. 4. 4 The Mercari app is a C2C marketplace where individuals can easily sell used items What is Mercari? Japan U.S. Monthly active users: 16+ Million Total number of items: 1.5+ Billion
  5. 5. 5 Why Content Moderation in C2C e-Commerce? C2C e-Commerce Sellers Buyers We want to decrease risk for customer and marketplace Sellers unintentionally violate policy. Buyers buy violated items without knowing Policy case: counterfeits, weapons, etc.
  6. 6. 6 Content Moderation system C2C e-Commerce Sell items Discover Moderator Manual review Moderation Service Hide items & Alert marketplace Sellers Buyers screened
  7. 7. 7 Concept of Moderation Service: Rule based Moderation Service Rule based Pros ● Easy to develop and can be quickly released to production Cons ● Hard to manage ● Difficult to cover the inconsistencies in spellings e.g. {NIKE, nike, ないき, ナイキ} Moderator Manual review
  8. 8. 8 Concept of Moderation Service: ML Moderation Service Rule based Pros ● Automatically learns the features of items deleted by moderators ● Adapts to spelling inconsistencies Cons ● Model update is hard ● Concept drift (a.k.a. training-serving skew) Moderator Manual review Machine Learning
  9. 9. 9 How to create the data for ML Rule based Moderator Machine Learning Sell items Report items Hide items & Alert Positive Deleted items by Moderator Negative Not deleted items by Moderator Dataset Moderation Service Review
  10. 10. 10 Task Design ● Data is highly imbalanced ● Each violated topic’s total number of alerts is bounded by moderator team All models trained as one-vs-all ● No side-effect when deploying a trained model to other class ● Hard to improve performance for each topic in a multi-class model Negative Violated Topic A Violated Topic N ... Positive Model A Model B ... counterfeits weapons
  11. 11. 11 Multimodality of content Case of items Items have multimodal data ● Image ● Text ● Category ● Brand ● Price, etc. We use multimodal model to improve model performance. See our article: https://tech.mercari.com/entry/2019/09/12/130000
  12. 12. 12 Model selection based on dataset size ● Gradient Boosted Decision Trees (GBDT) → Efficient for training and inference when training data size is not large *Image feature is not used in GBDT ● Gated Multimodal Unit (GMU) → Potentially most accurate using multimodal data
  13. 13. 13 Offline evaluation Metric is Precision@K: K is the bound on the daily total number of alerts in each violated topic decided by Moderators 2020-07-13 Current model’s prediction result In production Top K Evaluate new model against current model using the same item ids item ids same as production top K 2020-07-13 New model’s prediction result In test dataset. e.g.
  14. 14. 14 Online evaluation → Faster decision making leads to efficient operation Current Model New Model Same traffic Moderator Manual review Each model alert number: K/2 Metrics: Precision@K/2 After a certain time after a new model is released, we decide which model should be deprecated based on the above metrics. Classic A/B testing can take several months. It was difficult to collect enough transactions for t-test.
  15. 15. 15 Offline/online evaluation result Algorithms Offline Online GBDT +18.2% Not Released GMU +21.2% +23.2% Table shows the relative performance gain of offline evaluation metric is precision@K , online evaluation metric is precision@K/2 on one violated topic Baseline model is Logistic regression that was already released in production
  16. 16. 16 Container based Training Pipeline Data Load Write manifest files containing requirements like CPU, GPU and Storage CPU CPU or GPU Training Offline Evaluation CPU BigQueryBigQuery
  17. 17. 17 Serving system architecture Message queue Message queue proxy layer prediction layer . Preprocessing + inference Container Pod GBDT based model Preprocessing Container ..Proxy container subscribe publish Pod Inference Container Caffe2 Pod Deep Learning based model We manage over 15 Machine Learning models in production Pod Deep Learning based model
  18. 18. 18 Horizontal Pod Autoscaler by kubernetes ● Reliable system: Traffic changes with time, HPA can adopt to varying traffic ● Cheaper billing cost: Reduce to 1/6 by HPA Billing cost transition after applying HPA Billing cost day Each color is each machine learning model
  19. 19. 19 Impact of Machine Learning system Discovered 100 violating items Moderator Manual review Moderation Service Rule based Machine Learning Hide & Alert +Discovered 554 violating items Machine Learning system has increased coverage by 554% ↑ over rule based approach e.g.
  20. 20. 20 If you have a question to this talk First author is Shunya UETA, please e-mail: hurutoriya@mercari.com Acknowledgements Co-Authors: Suganprabu Nagarajan, Mizuki Sango Contributter: ● Abhishek Vilas Munagekar, Yusuke Shido, Vamshi Teja Racha, Sumit Verma and Keisuke Umezawa for their contribute to this system ● Dr. Antony for his feedback about the paper ● Yushi Kurita, Yuki Ito as Product Manager, All Trust and Safety project member and all Customer Service as Moderator to success this project. Question and Thanks collaborator
  • shunyaueta

    Jul. 29, 2020

2020 USENIX Conference on Operational Machine Learning Authors: Shunya Ueta, Suganprabu Nagaraja, and Mizuki Sango, Mercari, inc. Abstract: Consumer-to-consumer (C2C) e-Commerce is a large and growing industry with millions of monthly active users. In this paper, we propose auto content moderation for C2C e-Commerce to moderate items using Machine Learning (ML). We will also discuss practical knowledge gained from our auto content moderation system. The system has been deployed to production at Mercari since late 2017 and has significantly reduced the operation cost in detecting items violating our policies. This system has increased coverage by 554.8 % over a rule-based approach.

Views

Total views

796

On Slideshare

0

From embeds

0

Number of embeds

264

Actions

Downloads

7

Shares

0

Comments

0

Likes

1

×