MOV の機械学習システムを支える MLOps 実践【DeNA TechCon 2020 ライブ配信】

Mar 16, 20206 likes15,214 views

本セッションでは機械学習システム特有の課題を踏まえ、『お客さま探索ナビ』の開発・運用における MLOps の実践について、実際の取り組みとともに紹介します。具体的には、MLOps 技術の選定理由とナレッジ、MLOps のハマりポイントと解決策、さらなる自動化に向けて現在取り組んでいる内容などをお話しします。

DeNA TechCon 2020
#denatechcon
々
Map data ©2019 Google

DeNA TechCon 2020
#denatechcon
々
AI Platform TrainingBigQuery

DeNA TechCon 2020
#denatechcon
々
々
々
々々々

DeNA TechCon 2020
#denatechcon
def task_fail_slack_alert(context):
slack_msg = “Task Failed.”
failed_alert = SlackWebhookOperator(
task_id="xxx",
http_conn_id="slack",
webhook_token=Variable.get("webhook.token"),
message=slack_msg,
username="airflow",
)
return failed_alert.execute(context=context)

DeNA TechCon 2020
#denatechcon
<!subteam^ABCDEF123|@user_group>

DeNA TechCon 2020
#denatechcon
※
@pytest.fixture(scope="session", autouse=True)
def setup_teardown(request, test_config):
# setup
#
suffix = uuid.uuid4()
create_dataset(suffix)
yield
# teardown
#
delete_blobs(suffix)
delete_dataset(suffix)

DeNA TechCon 2020
#denatechcon
@pytest.mark.dependency(depends=[“test_preprocess_a”, “test_preprocess_b”])
def test_trainer(test_config):
trainer = Trainer(
test_config[“xxx”].format(test_config[“hash"]),
test_config[“yyy”].format(test_config[“hash"]),
)
trainer.train()
trainer.upload_model()

SSII2020 [OS2-02] 教師あり事前学習を凌駕する「弱」教師あり事前学習 6月11日 (木) 11:00 - 12:30 メイン会場（vimeo + slido）登壇者：片岡裕雄氏（産業技術総合研究所）概要：高品質な教師ラベルを含む大規模画像データセットを学習に用いた場合は成功がほぼ確約されていると言っても差し支えないが、ラベル付けのコストは計り知れない。本発表では、SNS からの画像データ収集や弱教師によるラベル付けを含め、最小限の手間で大規模データセットを構築する方法を紹介する。800万超の人物ラベルを含む大規模データセットを人物検出のための事前学習に用いた場合、ベースラインである教師ありの事前学習手法を凌駕する検出率を実現した。

Kubernetesによる機械学習基盤への挑戦Preferred Networks

2018年12月4日　Japan Container Days　講演資料谷脇大輔 Preferred Networksでは1000個以上のGPUとInfiniBandからなるオンプレミスのクラスターを自社で構築しており、研究者が様々な目的、リソース量、実行時間の機械学習ジョブをKubernetes上で実行して研究成果を出しています。 KubernetesはKubeflowの登場など、機械学習基盤としても非常に注目されている一方で、現実的には未だ発展途上です。講演では機械学習基盤としてのKubernetesの導入理由、その実用性と将来性、Preferred Networksの挑戦についてお話ししました。

[DL輪読会]Transframer: Arbitrary Frame Prediction with Generative ModelsDeep Learning JP

[DL輪読会]Pay Attention to MLPs （gMLP）Deep Learning JP

The document summarizes a research paper that compares the performance of MLP-based models to Transformer-based models on various natural language processing and computer vision tasks. The key points are: 1. Gated MLP (gMLP) architectures can achieve performance comparable to Transformers on most tasks, demonstrating that attention mechanisms may not be strictly necessary. 2. However, attention still provides benefits for some NLP tasks, as models combining gMLP and attention outperformed pure gMLP models on certain benchmarks. 3. For computer vision, gMLP achieved results close to Vision Transformers and CNNs on image classification, indicating gMLP can match their data efficiency.

【DL輪読会】Egocentric Video Task Translation (CVPR 2023 Highlight)Deep Learning JP

【DL輪読会】時系列予測 Transfomers の精度向上手法Deep Learning JP

Teslaにおけるコンピュータビジョン技術の調査Kazuyuki Miyazawa

【DL輪読会】Flamingo: a Visual Language Model for Few-Shot Learning 画像×言語の大規模基盤モ...Deep Learning JP

[DL輪読会]Grokking: Generalization Beyond Overfitting on Small Algorithmic DatasetsDeep Learning JP

Deep Learning による視覚×言語融合の最前線Yoshitaka Ushiku

2016/12/17 第7回 Machine Learning 15minutes!＠メンバーズ 2017/03/12 ステアラボ人工知能シンポジウム2017@AP秋葉原 2017/03/22 電子情報通信学会総合大会企画講演セッション「もっと知りたい！ Deep Learning 〜基礎から活用ノウハウ，応用まで〜」@名城大学 2017/06/30 ABEJA Technopreneur College にて一部もしくは全スライドを使用。画像キャプション生成については (https://www.slideshare.net/YoshitakaUshiku/ss-57148161) により詳細な説明を譲りますが、画像×言語の研究に関する日本語資料としての網羅的をより高めるように試みた資料です。

Data-Centric AIの紹介Kazuyuki Miyazawa

深層強化学習による自動運転車両の経路探索に関する研究harmonylab

近年，情報通信技術の発展に伴ってリアルタイムな交通状況を容易に入手できるようになってきている．そして，交通状況に基づいて目的地まで円滑に移動可能な経路を提供するサービスも普及し始めている．しかし，実際の交通状況においては現在の交通状況に関する情報だけで経路を考えるのでは不十分で，今後の交通流を予測する必要がある．本研究では，時間変化する交通状況において目的地まで短い時間で到達できる経路探索を実現することを目的として，深層強化学習を用いた経路探索手法の検証を行った．

ChatGPTは思ったほど賢くないCarnot Inc.

Teslaにおけるコンピュータビジョン技術の調査 (2)Kazuyuki Miyazawa

深層学習の不確実性 - Uncertainty in Deep Neural Networks -tmtm otm

【DL輪読会】ConvNeXt V2: Co-designing and Scaling ConvNets with Masked AutoencodersDeep Learning JP

PFNのML/DL基盤を支えるKubernetesにおける自動化 / DevOpsDays Tokyo 2021Preferred Networks

Preferred Networks（PFN）は深層学習などの最先端の技術を最短路で実用化することで、これまで解決が困難であった現実世界の課題解決を目指しています。コンピュータビジョン、自然言語処理、音声認識、ロボティクス、コンパイラ、分散処理、専用ハードウェア、バイオインフォマティクス、ケモインフォマティクスといった幅広い分野で研究開発を行っており、それを支えているのが Kubernetes を用いて構築しているオンプレミス/ベアメタルの GPU クラスタです。本セッションでは、PFN が Kubernetes を用いてクラスタを運用するなかでどのような障害が起きるのかを紹介し、また障害対応をどのように自動化しているのかを具体的に使用/開発したソフトウェアを含めてご紹介します。また Kubernetes クラスタの管理、アップグレードの自動化にも取り組んでおり、それを実現する Cluster API についてもご紹介します。 https://confengine.com/conferences/devopsdays-tokyo-2021/proposal/15203/pfn-mldl-kubernetes

Action Recognitionの歴史と最新動向Ohnishi Katsunori

This document summarizes recent developments in action recognition using deep learning techniques. It discusses early approaches using improved dense trajectories and two-stream convolutional neural networks. It then focuses on advances using 3D convolutional networks, enabled by large video datasets like Kinetics. State-of-the-art results are achieved using inflated 3D convolutional networks and temporal aggregation methods like temporal linear encoding. The document provides an overview of popular datasets and challenges and concludes with tips on training models at scale.

ナレッジグラフ入門KnowledgeGraph

プレゼン・ポスターで自分の研究を「伝える」　(How to do technical oral/poster presentation)Toshihiko Yamasaki

MIRU2020若手プログラム招待講演のスライドを一般公開用にアレンジしたものです。日本語で書かれています。下記の点にご注意ください・セリフが伴ってないので内容は限定的です・著作権等に配慮しているので中身は結構無味乾燥です。 This is an arranged version of my invited talk at MIRU 2020 young researchers' forum. This is written in Japanese.

Swin Transformer (ICCV'21 Best Paper) を完璧に理解する資料Yusuke Uchida

最近のディープラーニングのトレンド紹介_20200925小川雄太郎

・自己教師あり学習・半教師あり学習・転移学習・メタ学習・模倣学習・逆強化学習・能動学習（Active Learning）について「概要」、「代表的アルゴリズム名」、「各内容を学ぶためのおすすめチュートリアル動画」を解説しています 2020年9月25日理系ナビ　DLエンジニアトレーニングプログラム

動作認識の最前線：手法，タスク，データセットToru Tamaki

ゼロから始める転移学習Yahoo!デベロッパーネットワーク

MLOps に基づく AI/ML 実運用最前線～画像、動画データにおける MLOps 事例のご紹介～（映像情報メディア学会2021年冬季大会企画セッショ...NTT DATA Technology & Innovation

[DL輪読会]Focal Loss for Dense Object DetectionDeep Learning JP

トピックモデルの評価指標 Perplexity とは何なのか？hoxo_m

Physics-ML のためのフレームワーク NVIDIA Modulus 最新事情NVIDIA Japan

仕様起因の手戻りを減らして開発効率アップを目指すチャレンジ【DeNA TechCon 2020 ライブ配信】DeNA

Flutter Vikings 2022 - End to end IoT with Dart and FlutterChris Swan

Things need apps to manage them, which Flutter is great for, providing an easy way to build cross platform support. But things also need to get their data (securely and privately) to their apps, and Dart can be used for that. This presentation will walk through a use case demonstrated at Mobile World Congress (and now open sourced) that uses Dart to read sensor data through to Flutter for user presentation.

More Related Content

What's hot (20)

[DL輪読会]Grokking: Generalization Beyond Overfitting on Small Algorithmic DatasetsDeep Learning JP

Deep Learning による視覚×言語融合の最前線Yoshitaka Ushiku

Data-Centric AIの紹介Kazuyuki Miyazawa

深層強化学習による自動運転車両の経路探索に関する研究harmonylab

ChatGPTは思ったほど賢くないCarnot Inc.

Teslaにおけるコンピュータビジョン技術の調査 (2)Kazuyuki Miyazawa

深層学習の不確実性 - Uncertainty in Deep Neural Networks -tmtm otm

【DL輪読会】ConvNeXt V2: Co-designing and Scaling ConvNets with Masked AutoencodersDeep Learning JP

PFNのML/DL基盤を支えるKubernetesにおける自動化 / DevOpsDays Tokyo 2021Preferred Networks

Action Recognitionの歴史と最新動向Ohnishi Katsunori

ナレッジグラフ入門KnowledgeGraph

プレゼン・ポスターで自分の研究を「伝える」　(How to do technical oral/poster presentation)Toshihiko Yamasaki

Swin Transformer (ICCV'21 Best Paper) を完璧に理解する資料Yusuke Uchida

最近のディープラーニングのトレンド紹介_20200925小川雄太郎

動作認識の最前線：手法，タスク，データセットToru Tamaki

ゼロから始める転移学習Yahoo!デベロッパーネットワーク

[DL輪読会]Focal Loss for Dense Object DetectionDeep Learning JP

トピックモデルの評価指標 Perplexity とは何なのか？hoxo_m

Physics-ML のためのフレームワーク NVIDIA Modulus 最新事情NVIDIA Japan

[DL輪読会]Grokking: Generalization Beyond Overfitting on Small Algorithmic DatasetsDeep Learning JP

Deep Learning による視覚×言語融合の最前線Yoshitaka Ushiku

Data-Centric AIの紹介Kazuyuki Miyazawa

深層強化学習による自動運転車両の経路探索に関する研究harmonylab

ChatGPTは思ったほど賢くないCarnot Inc.

Teslaにおけるコンピュータビジョン技術の調査 (2)Kazuyuki Miyazawa

深層学習の不確実性 - Uncertainty in Deep Neural Networks -tmtm otm

【DL輪読会】ConvNeXt V2: Co-designing and Scaling ConvNets with Masked AutoencodersDeep Learning JP

PFNのML/DL基盤を支えるKubernetesにおける自動化 / DevOpsDays Tokyo 2021Preferred Networks

Action Recognitionの歴史と最新動向Ohnishi Katsunori

ナレッジグラフ入門KnowledgeGraph

プレゼン・ポスターで自分の研究を「伝える」　(How to do technical oral/poster presentation)Toshihiko Yamasaki

Swin Transformer (ICCV'21 Best Paper) を完璧に理解する資料Yusuke Uchida

最近のディープラーニングのトレンド紹介_20200925小川雄太郎

動作認識の最前線：手法，タスク，データセットToru Tamaki

ゼロから始める転移学習Yahoo!デベロッパーネットワーク

[DL輪読会]Focal Loss for Dense Object DetectionDeep Learning JP

トピックモデルの評価指標 Perplexity とは何なのか？hoxo_m

Physics-ML のためのフレームワーク NVIDIA Modulus 最新事情NVIDIA Japan

Similar to MOV の機械学習システムを支える MLOps 実践【DeNA TechCon 2020 ライブ配信】 (20)

仕様起因の手戻りを減らして開発効率アップを目指すチャレンジ【DeNA TechCon 2020 ライブ配信】DeNA

Flutter Vikings 2022 - End to end IoT with Dart and FlutterChris Swan

UK WVD User Group January 2021 SlidesNeil McLoughlin

The document summarizes the agenda for the UK WVD User Group meeting in January 2021. It includes introductions, what's new with WVD this month, personal news from Neil McLoughlin about joining Nerdio as their UK Field CTO, and a presentation on monitoring, managing and optimizing WVD with ControlUp. There will also be a presentation on designing and implementing BC/DR in a WVD environment, followed by a Q&A session.

Roche | The Design History File in codeBeamer ALM: Electronic Records, Signat...Intland Software GmbH

IoT with circuitpython | CloudBadgeAyan Pahwa

The document discusses the development of an Internet of Things (IoT) conference badge called the PyBadge. It outlines the key components needed for a connected conference badge including hardware, software, cloud connectivity, and communication protocols. It then demonstrates a prototype PyBadge that uses a microcontroller, display, and WiFi to connect to a cloud-based dashboard for controlling the badge's LED lights and name over MQTT. Future work is mentioned to add Bluetooth support, gaming emulators, and communication capabilities to the PyBadge design.

Fairfax Sydney #mojo #mojocon Feb 2017 Meetup - 360 video production basicsJamie Andrei

The future of WebRTC - Sept 2021Arnaud BUDKIEWICZ

Tekton showcase - CDF Summit Kubecon Barcelona 2019Christie Wilson

Tekton is an open source project that provides containers-native pipeline functionality for Kubernetes. The presentation introduced Tekton Pipelines and its components like Tasks and Pipelines custom resources. Demos were shown of Tekton Pipelines, the tkn CLI tool, and the Tekton Dashboard UI. Integration patterns for using Tekton with other CI/CD tools were discussed. The roadmap outlines completing Tekton Pipelines in 2019 with features like event triggering and SCM support.

Sony Vegas Pro 13.0 Instruction Manual and Guide - EnglishRahul Jose

OSMC 2019 | The Telegraf Toolbelt: It Can Do That, Really? by David McKayNETWAYS

Telegraf is an agent for collecting, processing, aggregating, and writing metrics. With over 200 plugins, Telegraf can fetch metrics from a variety of sources, allowing you to build aggregations and write those metrics to InfluxDB, Prometheus, Kafka, and more. In this talk, we will take a look at some of the lesser known, but awesome, plugins that are often overlooked; as well as how to use Telegraf for monitoring of Cloud Native systems.

WebRTC: The Big Debate, Shut Up and Build SomethingDigium

The document discusses WebRTC, a technology that enables real-time communication via video chat and file sharing directly in web browsers. It provides an overview of WebRTC's history and key features, discusses popular use cases and production applications using WebRTC like Amazon Mayday and Comcast X1. The presentation encourages attendees to build solutions with WebRTC rather than debate its technical aspects and details recent acquisitions by companies wanting to integrate WebRTC into their products.

Astricon 2014 - WebRTC - The Big Debate, I Say Shut Up and Build Something - ...Blacc Spot Media, Inc.

Lantre Barr presented on WebRTC and discussed its history, key features, use cases, and production applications. Some of the applications discussed include Amazon Mayday Button, Comcast X1 Platform, and Esurance Video Appraisal. Barr emphasized building with WebRTC rather than debating details, and discussed acquisitions like Snapchat acquiring AddLive to utilize WebRTC. Barr concluded by sharing free WebRTC services and resources for learning more.

Flutter Festival London 2022 - End to end IoT with Dart and FlutterChris Swan

Back-2-Basics: Code ContractsDavid McCarter

This document summarizes a presentation on code contracts, which are preconditions, postconditions, and class invariants that define the expectations and promises of software elements. The presentation discusses the background of design by contract and how contracts can be used throughout the software lifecycle to write better code, analyze designs, implement and reuse components properly, avoid bugs, automatically document software, help with testing and debugging, and assist project managers. It also advertises a DVD and conference photos for sale, and lists local user groups in San Diego.

WaveEngine Dotnet 2018Javier Cantón Ferrero

The document summarizes Javier Canton's presentation at #DotNet2018 about developing a 3D graphics engine in C# for virtual reality and augmented reality. It provides an overview of Javier's background and experience developing 3D graphics. It also summarizes his demonstration of the Wave Engine, a C# graphics engine he developed that supports VR, AR and MR across platforms. The presentation covered various C# language features relevant for graphics development and performance optimization.

LightscribeNathan Wade

This document discusses different methods for labeling discs, including hand labeling, paper labels, printable discs, professional silkscreening, and Lightscribe labeling. It provides pros and cons of each method. Lightscribe allows labeling discs using a laser inside a Lightscribe-enabled disc drive without ink, is low cost, and provides a professional look. The document outlines the Lightscribe labeling process and provides steps for making an audio CD label using Lightscribe.

Graphics of the Americas 2017 Attendee BrochureDepo Consulting

This document provides information about the GOAexpo trade show that will take place from February 16-18, 2017 at the Broward County Convention Center. The trade show will feature over 40 seminars on topics like Adobe Creative Cloud, photography, prepress, production, leadership, and wide format printing. It will also include exhibit areas for wide format printing, digital flexo printing, dye sublimation, 3D printing, and a career fair. The document encourages graphic arts professionals to attend in order to learn new skills, see the latest equipment and services, and network with industry experts.

FIWARE Global Summit - Real-time Media Stream Processing using KurentoFIWARE

When we design togetherThoughtworks

When we design together - Sabrina Mach, Ammara Gafoor and James Emmott From three distinct perspectives, this talk will contend that design is an activity undertaken by everyone in a software development team. It occurs throughout the process of delivery — not only at the beginning or the end — and it is a powerful instrument for learning about and adapting to the problems our work seeks to solve, which is a shared responsibility. Making the best use of our multidisciplinary expertise in the activity of design requires forms of collaboration that are too often disrupted by the role-based silos that keep us separated and weaken the valuable contribution our diverse approaches could make to our collective efforts. If you care about accelerating time to market, improving customer experience, or building happy and productive teams, you will want to know why and how it matters that we believe ‘design is in everything that we do’.

Doodle maker review special offer - doodle maker featuresrajugond2

仕様起因の手戻りを減らして開発効率アップを目指すチャレンジ【DeNA TechCon 2020 ライブ配信】DeNA

Flutter Vikings 2022 - End to end IoT with Dart and FlutterChris Swan

UK WVD User Group January 2021 SlidesNeil McLoughlin

Roche | The Design History File in codeBeamer ALM: Electronic Records, Signat...Intland Software GmbH

IoT with circuitpython | CloudBadgeAyan Pahwa

Fairfax Sydney #mojo #mojocon Feb 2017 Meetup - 360 video production basicsJamie Andrei

The future of WebRTC - Sept 2021Arnaud BUDKIEWICZ

Tekton showcase - CDF Summit Kubecon Barcelona 2019Christie Wilson

Sony Vegas Pro 13.0 Instruction Manual and Guide - EnglishRahul Jose

OSMC 2019 | The Telegraf Toolbelt: It Can Do That, Really? by David McKayNETWAYS

WebRTC: The Big Debate, Shut Up and Build SomethingDigium

Astricon 2014 - WebRTC - The Big Debate, I Say Shut Up and Build Something - ...Blacc Spot Media, Inc.

Flutter Festival London 2022 - End to end IoT with Dart and FlutterChris Swan

Back-2-Basics: Code ContractsDavid McCarter

WaveEngine Dotnet 2018Javier Cantón Ferrero

LightscribeNathan Wade

Graphics of the Americas 2017 Attendee BrochureDepo Consulting

FIWARE Global Summit - Real-time Media Stream Processing using KurentoFIWARE

When we design togetherThoughtworks

Doodle maker review special offer - doodle maker featuresrajugond2

More from DeNA (20)

DRIVE CHARTの裏側〜 AI ☓ IoT ☓ ビッグデータを支えるアーキテクチャ〜DeNA

年間40万件以上発生している交通事故。DRIVE CHARTはAIドラレコで交通事故ゼロ社会を目指すサービスです。 DRIVE CHARTを実現する上では、AIの組み込みやIoT機器によるセンシング、膨大なデータ処理といったチャレンジがありました。このセッションではその裏側を支えるサーバサイドおよびIoT機器の事例についてご紹介します。サーバサイドにおいては、セキュアでスケーラブルなアーキテクチャの設計、多言語・多数のコンポーネント間のマイクロサービス化、IoT機器やAIシステムとの膨大な動画・データ処理連携における工夫、OpenAPIDocumentを利用したAPI自動テストの効率化をご紹介したいと思います。 IoT機器においては、リアルタイムでのAI活用で発生した課題とその対策・検証、IoT機器の移動体通信の仕組みと通信量削減の手法について具体的な事例をご紹介したいと思います。

IoTと業務システムをつなぐgRPC/RESTサービスの開発と運用DeNA

次世代タクシー配車アプリ「MOV」では、配車アプリだけでなくタクシー事業者様向けの機能開発も行っています。私たちのチームは、「MOV」システムで収集した車両データを利用し、従来のタクシー配車システムとタクシー車両の無線連携の仕組みをインターネット上で実現しようとしています。本セッションでは、GKE, gRPC, OpenAPIなどの技術を駆使して、タクシー配車システムに求められる大量の車両データのリアルタイム配信や、タクシー事業者様ごとに異なる配車システムの言語, プロトコルに向き合うためのシステム構築の工夫や運用について事例を交えてご紹介します。

Can We Make Maps from Videos? ~From AI Algorithm to Engineering for Continuou...DeNA

SHOWROOMとDeNAで取り組んだライブ配信基盤刷新・超低遅延ライブ配信の裏側【DeNA TechCon 2020 ライブ配信】DeNA

本セッションでは、以前のアーキテクチャと比較してどう変わったのか、変える上で重要視したポイントは何か、超低遅延配信の詳細、移行時の失敗談等について触れながら、今回のライブ配信基盤の抜本的なアーキテクチャ刷新の取り組みについてご紹介します。 #livestreaming #HLS #LHLS #RTMP #超低遅延 #AWS #IDCF #Wowza #CDN

クラウド環境でのセキュリティ監査自動化【DeNA TechCon 2020 ライブ配信】DeNA

このセッションでは、クラウド環境で多数のアカウントを利用した際に発生するセキュリティ的な問題と、それを解決するために作成した危険な設定を自動的にチェックするシステムを紹介します。今回作成したシステムと既製品との違いや具体的なチェック内容、システムを作った後にどのようにして社内の運用に乗せていったのかも解説します。

DeClang 誕生！Clang ベースのハッキング対策コンパイラ【DeNA TechCon 2020 ライブ配信】DeNA

DeNA データプラットフォームにおける自由と統制のバランス【DeNA TechCon 2020 ライブ配信】DeNA

リアルタイムリモートデバッグ環境によるゲーム開発イテレーションの高速化【DeNA TechCon 2020 ライブ配信】DeNA

本セッションでは、Unityを使った開発のイテレーション高速化を目的として、有線接続したモバイル実機のゲームに対して開発用PCからリアルタイムに介入できるツールを高い柔軟性で容易に開発できる基盤をgRPC等と独自TCPリレーサーバーを組み合わせて構築した事例と、その上に実現したファイル転送システムをはじめとした、具体的に構築したデバッグツールの事例をご紹介します。

コンピュータビジョン技術の実応用〜DRIVE CHARTにおける脇見・車間距離不足検知〜【DeNA TechCon 2020 ライブ配信】DeNA

The document appears to be slides from DeNA TechCon 2020. It discusses several topics relating to AI and computer vision, including DRIVE CHART which is DeNA's AI platform, work on object detection, optical flow estimation, and the use of neural networks like CNNs and RNNs in areas like computer vision. It also references prior work on generative adversarial networks and improving discriminators and generators in these networks.

DeNA の Slack 導入と活用の事例紹介DeNA

第2回ゲーム開発者向けイベント〜現場でのSlack活用術〜での登壇資料です。 https://peatix.com/event/1394472 システム本部/IT戦略部佐藤大輝・DeNAのSlack導入経緯とオフィスの不具合報告チャンネルゲーム・エンターテインメント事業本部ゲーム事業部越智琢正・頑張りすぎない活用事例: 人間と機械の連携ゲーム・エンターテインメント事業本部ゲーム事業部今江健悟・社内勉強会におけるSlack活用事例 ~Slackのコメントをニコニコ動画風に投影~

タクシーxAIを支えるKubernetesとAIデータパイプラインの信頼性の取り組みについて [SRE NEXT 2020]DeNA

AIを使ったオートモーティブ領域のサービスとして、タクシー営業を最適化するお客様探索ナビを開発しています。データサイエンティスト、ドメインアルゴリズムエンジニアが開発した結果を、MLOps、サーバサイドエンジニアがデータパイプラインとサーバAPIとして構築し、最終的にタクシー乗務員向けの最適走行ナビアプリとして提供しています。このアーキテクチャの信頼性に対する取り組みとして、データパイプライン中でタクシー運行シミュレーションによる評価行ったり、API末端で案内経路に異常が発生していないか、交通規制を守って走行できるかのチェックなど、複数の施策を行っています。スコープをアーキテクチャの信頼性にしぼり、データパイプラインとアーキテクチャの全体像をお話した後、信頼性を担保するために、行っている施策とそれを取り組むに至った考え方を紹介します。

オートモーティブ領域における位置情報関連アルゴリズムあれこれDeNA

後部座席タブレットにおけるMaaS時代を見据えた半歩先のUX設計」 [MOBILITY:dev]DeNA

ドライブレコーダ映像からの3次元空間認識 [MOBILITY:dev]DeNA

MOVで実践したサーバーAPI実装の超最適化について [MOBILITY:dev]DeNA

MOV お客さま探索ナビの GCP ML開発フローについてDeNA

課題ドリブン、フルスタックAI開発術 [MOBILITY:dev]DeNA

AIの精度向上において、学習データやモデルの出力の可視化や解析は非常に重要です。我々は、これまで手動で行っていたそれらのプロセスをマイクロサービスに置き換え自動化することで大幅な効率化を実現しました。本セッションでは、そうしたフルスタックなAI開発術についてお話しします。

DeNA の AWS アカウント管理とセキュリティ監査自動化DeNA

AWSの利用で以下の課題はありませんか？・多数のAWSアカウントがあり管理が難しい・セキュリティ標準がなく設定者により差異がでる・セキュリティ監査工数が膨大になるこれらの対策として・多数のAWSアカウント管理を効率化した方法・AWS利用のセキュリティ標準策定や監査の自動化などをご紹介した資料になります。

DeNAのQCTマネジメント IaaS利用のベストプラクティス [AWS Summit Tokyo 2019]DeNA

知っててもいいかもしれない知財のこと（抜粋版）DeNA

DRIVE CHARTの裏側〜 AI ☓ IoT ☓ ビッグデータを支えるアーキテクチャ〜DeNA

IoTと業務システムをつなぐgRPC/RESTサービスの開発と運用DeNA

Can We Make Maps from Videos? ~From AI Algorithm to Engineering for Continuou...DeNA

SHOWROOMとDeNAで取り組んだライブ配信基盤刷新・超低遅延ライブ配信の裏側【DeNA TechCon 2020 ライブ配信】DeNA

クラウド環境でのセキュリティ監査自動化【DeNA TechCon 2020 ライブ配信】DeNA

DeClang 誕生！Clang ベースのハッキング対策コンパイラ【DeNA TechCon 2020 ライブ配信】DeNA

DeNA データプラットフォームにおける自由と統制のバランス【DeNA TechCon 2020 ライブ配信】DeNA

リアルタイムリモートデバッグ環境によるゲーム開発イテレーションの高速化【DeNA TechCon 2020 ライブ配信】DeNA

コンピュータビジョン技術の実応用〜DRIVE CHARTにおける脇見・車間距離不足検知〜【DeNA TechCon 2020 ライブ配信】DeNA

DeNA の Slack 導入と活用の事例紹介DeNA

タクシーxAIを支えるKubernetesとAIデータパイプラインの信頼性の取り組みについて [SRE NEXT 2020]DeNA

オートモーティブ領域における位置情報関連アルゴリズムあれこれDeNA

後部座席タブレットにおけるMaaS時代を見据えた半歩先のUX設計」 [MOBILITY:dev]DeNA

ドライブレコーダ映像からの3次元空間認識 [MOBILITY:dev]DeNA

MOVで実践したサーバーAPI実装の超最適化について [MOBILITY:dev]DeNA

MOV お客さま探索ナビの GCP ML開発フローについてDeNA

課題ドリブン、フルスタックAI開発術 [MOBILITY:dev]DeNA

DeNA の AWS アカウント管理とセキュリティ監査自動化DeNA

DeNAのQCTマネジメント IaaS利用のベストプラクティス [AWS Summit Tokyo 2019]DeNA

知っててもいいかもしれない知財のこと（抜粋版）DeNA

Recently uploaded (20)

Cyber Security Legal Framework in Nepal.pptxGhimire B.R.

Cyber security cyber security cyber security cyber security cyber security cy...pranavbodhak

Gihbli AI and Geo sitution |use/misuse of Ai Technologyzainkhurram1111

End-to-end Assurance for SD-WAN & SASE with ThousandEyesThousandEyes

Let’s Get Slack Certified! 🚀- Slack CommunitySanjeetMishra29

Microsoft Build 2025 takeaways in one presentationDigitalmara

Microsoft Build 2025 introduced significant updates. Everything revolves around AI. DigitalMara analyzed these announcements: • AI enhancements for Windows 11 By embedding AI capabilities directly into the OS, Microsoft is lowering the barrier for users to benefit from intelligent automation without requiring third-party tools. It's a practical step toward improving user experience, such as streamlining workflows and enhancing productivity. However, attention should be paid to data privacy, user control, and transparency of AI behavior. The implementation policy should be clear and ethical. • GitHub Copilot coding agent The introduction of coding agents is a meaningful step in everyday AI assistance. However, it still brings challenges. Some people compare agents with junior developers. They noted that while the agent can handle certain tasks, it often requires supervision and can introduce new issues. This innovation holds both potential and limitations. Balancing automation with human oversight is crucial to ensure quality and reliability. • Introduction of Natural Language Web NLWeb is a significant step toward a more natural and intuitive web experience. It can help users access content more easily and reduce reliance on traditional navigation. The open-source foundation provides developers with the flexibility to implement AI-driven interactions without rebuilding their existing platforms. NLWeb is a promising level of web interaction that complements, rather than replaces, well-designed UI. • Introduction of Model Context Protocol MCP provides a standardized method for connecting AI models with diverse tools and data sources. This approach simplifies the development of AI-driven applications, enhancing efficiency and scalability. Its open-source nature encourages broader adoption and collaboration within the developer community. Nevertheless, MCP can face challenges in compatibility across vendors and security in context sharing. Clear guidelines are crucial. • Windows Subsystem for Linux is open-sourced It's a positive step toward greater transparency and collaboration in the developer ecosystem. The community can now contribute to its evolution, helping identify issues and expand functionality faster. However, open-source software in a core system also introduces concerns around security, code quality management, and long-term maintenance. Microsoft’s continued involvement will be key to ensuring WSL remains stable and secure. • Azure AI Foundry platform hosts Grok 3 AI models Adding new models is a valuable expansion of AI development resources available at Azure. This provides developers with more flexibility in choosing language models that suit a range of application sizes and needs. Hosting on Azure makes access and integration easier when using Microsoft infrastructure.

Droidal: AI Agents Revolutionizing HealthcareDroidal LLC

Droidal’s AI Agents are transforming healthcare by bringing intelligence, speed, and efficiency to key areas such as Revenue Cycle Management (RCM), clinical operations, and patient engagement. Built specifically for the needs of U.S. hospitals and clinics, Droidal's solutions are designed to improve outcomes and reduce administrative burden. Through simple visuals and clear examples, the presentation explains how AI Agents can support medical coding, streamline claims processing, manage denials, ensure compliance, and enhance communication between providers and patients. By integrating seamlessly with existing systems, these agents act as digital coworkers that deliver faster reimbursements, reduce errors, and enable teams to focus more on patient care. Droidal's AI technology is more than just automation — it's a shift toward intelligent healthcare operations that are scalable, secure, and cost-effective. The presentation also offers insights into future developments in AI-driven healthcare, including how continuous learning and agent autonomy will redefine daily workflows. Whether you're a healthcare administrator, a tech leader, or a provider looking for smarter solutions, this presentation offers a compelling overview of how Droidal’s AI Agents can help your organization achieve operational excellence and better patient outcomes. A free demo trial is available for those interested in experiencing Droidal’s AI Agents firsthand. Our team will walk you through a live demo tailored to your specific workflows, helping you understand the immediate value and long-term impact of adopting AI in your healthcare environment. To request a free trial or learn more: https://droidal.com/

Nix(OS) for Python Developers - PyCon 25 (Bologna, Italia)Peter Bittner

How do you onboard new colleagues in 2025? How long does it take? Would you love a standardized setup under version control that everyone can customize for themselves? A stable desktop setup, reinstalled in just minutes. It can be done. This talk was given in Italian, 29 May 2025, at PyCon 25, Bologna, Italy. All slides are provided in English. Original slides at https://slides.com/bittner/pycon25-nixos-for-python-developers

Improving Developer Productivity With DORA, SPACE, and DevExJustin Reock

Protecting Your Sensitive Data with Microsoft Purview - IRMS 2025Nikki Chapple

Session | Protecting Your Sensitive Data with Microsoft Purview: Practical Information Protection and DLP Strategies Presenter | Nikki Chapple (MVP| Principal Cloud Architect CloudWay) & Ryan John Murphy (Microsoft) Event | IRMS Conference 2025 Format | Birmingham UK Date | 18-20 May 2025 In this closing keynote session from the IRMS Conference 2025, Nikki Chapple and Ryan John Murphy deliver a compelling and practical guide to data protection, compliance, and information governance using Microsoft Purview. As organizations generate over 2 billion pieces of content daily in Microsoft 365, the need for robust data classification, sensitivity labeling, and Data Loss Prevention (DLP) has never been more urgent. This session addresses the growing challenge of managing unstructured data, with 73% of sensitive content remaining undiscovered and unclassified. Using a mountaineering metaphor, the speakers introduce the “Secure by Default” blueprint—a four-phase maturity model designed to help organizations scale their data security journey with confidence, clarity, and control. 🔐 Key Topics and Microsoft 365 Security Features Covered: Microsoft Purview Information Protection and DLP Sensitivity labels, auto-labeling, and adaptive protection Data discovery, classification, and content labeling DLP for both labeled and unlabeled content SharePoint Advanced Management for workspace governance Microsoft 365 compliance center best practices Real-world case study: reducing 42 sensitivity labels to 4 parent labels Empowering users through training, change management, and adoption strategies 🧭 The Secure by Default Path – Microsoft Purview Maturity Model: Foundational – Apply default sensitivity labels at content creation; train users to manage exceptions; implement DLP for labeled content. Managed – Focus on crown jewel data; use client-side auto-labeling; apply DLP to unlabeled content; enable adaptive protection. Optimized – Auto-label historical content; simulate and test policies; use advanced classifiers to identify sensitive data at scale. Strategic – Conduct operational reviews; identify new labeling scenarios; implement workspace governance using SharePoint Advanced Management. 🎒 Top Takeaways for Information Management Professionals: Start secure. Stay protected. Expand with purpose. Simplify your sensitivity label taxonomy for better adoption. Train your users—they are your first line of defense. Don’t wait for perfection—start small and iterate fast. Align your data protection strategy with business goals and regulatory requirements. 💡 Who Should Watch This Presentation? This session is ideal for compliance officers, IT administrators, records managers, data protection officers (DPOs), security architects, and Microsoft 365 governance leads. Whether you're in the public sector, financial services, healthcare, or education. 🔗 Read the blog: https://nikkichapple.com/irms-conference-2025/

TrustArc Webinar: Mastering Privacy ContractingTrustArc

As data privacy regulations become more pervasive across the globe and organizations increasingly handle and transfer (including across borders) meaningful volumes of personal and confidential information, the need for robust contracts to be in place is more important than ever. This webinar will provide a deep dive into privacy contracting, covering essential terms and concepts, negotiation strategies, and key practices for managing data privacy risks. Whether you're in legal, privacy, security, compliance, GRC, procurement, or otherwise, this session will include actionable insights and practical strategies to help you enhance your agreements, reduce risk, and enable your business to move fast while protecting itself. This webinar will review key aspects and considerations in privacy contracting, including: - Data processing addenda, cross-border transfer terms including EU Model Clauses/Standard Contractual Clauses, etc. - Certain legally-required provisions (as well as how to ensure compliance with those provisions) - Negotiation tactics and common issues - Recent lessons from recent regulatory actions and disputes

Cognitive Chasms - A Typology of GenAI Failure Failure ModesDr. Tathagat Varma

Cybersecurity Fundamentals: Apprentice - Palo Alto CertificateVICTOR MAESTRE RAMIREZ

STKI Israel Market Study 2025 final v1 versionDr. Jimmy Schwarzkopf

Jira Administration Training – Day 1 : IntroductionRavi Teja

Dr Jimmy Schwarzkopf presentation on the SUMMIT 2025 ADr. Jimmy Schwarzkopf

Introduction and Background: Study Overview and Methodology: The study analyzes the IT market in Israel, covering over 160 markets and 760 companies/products/services. It includes vendor rankings, IT budgets, and trends from 2025-2029. Vendors participate in detailed briefings and surveys. Vendor Listings: The presentation lists numerous vendors across various pages, detailing their names and services. These vendors are ranked based on their participation and market presence. Market Insights and Trends: Key insights include IT market forecasts, economic factors affecting IT budgets, and the impact of AI on enterprise IT. The study highlights the importance of AI integration and the concept of creative destruction. Agentic AI and Future Predictions: Agentic AI is expected to transform human-agent collaboration, with AI systems understanding context and orchestrating complex processes. Future predictions include AI's role in shopping and enterprise IT.

Offshore IT Support: Balancing In-House and Offshore Help Desk Techniciansjohn823664

In today's always-on digital environment, businesses must deliver seamless IT support across time zones, devices, and departments. This SlideShare explores how companies can strategically combine in-house expertise with offshore talent to build a high-performing, cost-efficient help desk operation. From the benefits and challenges of offshore support to practical models for integrating global teams, this presentation offers insights, real-world examples, and key metrics for success. Whether you're scaling a startup or optimizing enterprise support, discover how to balance cost, quality, and responsiveness with a hybrid IT support strategy. Perfect for IT managers, operations leads, and business owners considering global help desk solutions.

Measuring Microsoft 365 Copilot and Gen AI SuccessNikki Chapple

Session | Measuring Microsoft 365 Copilot and Gen AI Success with Viva Insights and Purview Presenter | Nikki Chapple 2 x MVP and Principal Cloud Architect at CloudWay Event | European Collaboration Conference 2025 Format | In person Germany Date | 28 May 2025 📊 Measuring Copilot and Gen AI Success with Viva Insights and Purview Presented by Nikki Chapple – Microsoft 365 MVP & Principal Cloud Architect, CloudWay How do you measure the success—and manage the risks—of Microsoft 365 Copilot and Generative AI (Gen AI)? In this ECS 2025 session, Microsoft MVP and Principal Cloud Architect Nikki Chapple explores how to go beyond basic usage metrics to gain full-spectrum visibility into AI adoption, business impact, user sentiment, and data security. 🎯 Key Topics Covered: Microsoft 365 Copilot usage and adoption metrics Viva Insights Copilot Analytics and Dashboard Microsoft Purview Data Security Posture Management (DSPM) for AI Measuring AI readiness, impact, and sentiment Identifying and mitigating risks from third-party Gen AI tools Shadow IT, oversharing, and compliance risks Microsoft 365 Admin Center reports and Copilot Readiness Power BI-based Copilot Business Impact Report (Preview) 📊 Why AI Measurement Matters: Without meaningful measurement, organizations risk operating in the dark—unable to prove ROI, identify friction points, or detect compliance violations. Nikki presents a unified framework combining quantitative metrics, qualitative insights, and risk monitoring to help organizations: Prove ROI on AI investments Drive responsible adoption Protect sensitive data Ensure compliance and governance 🔍 Tools and Reports Highlighted: Microsoft 365 Admin Center: Copilot Overview, Usage, Readiness, Agents, Chat, and Adoption Score Viva Insights Copilot Dashboard: Readiness, Adoption, Impact, Sentiment Copilot Business Impact Report: Power BI integration for business outcome mapping Microsoft Purview DSPM for AI: Discover and govern Copilot and third-party Gen AI usage 🔐 Security and Compliance Insights: Learn how to detect unsanctioned Gen AI tools like ChatGPT, Gemini, and Claude, track oversharing, and apply eDLP and Insider Risk Management (IRM) policies. Understand how to use Microsoft Purview—even without E5 Compliance—to monitor Copilot usage and protect sensitive data. 📈 Who Should Watch: This session is ideal for IT leaders, security professionals, compliance officers, and Microsoft 365 admins looking to: Maximize the value of Microsoft Copilot Build a secure, measurable AI strategy Align AI usage with business goals and compliance requirements 🔗 Read the blog https://nikkichapple.com/measuring-copilot-gen-ai/

Kubernetes Cloud Native Indonesia Meetup - May 2025Prasta Maha

GDG Cloud Southlake #43: Tommy Todd: The Quantum Apocalypse: A Looming Threat...James Anderson

The Quantum Apocalypse: A Looming Threat & The Need for Post-Quantum Encryption We explore the imminent risks posed by quantum computing to modern encryption standards and the urgent need for post-quantum cryptography (PQC). Bio: With 30 years in cybersecurity, including as a CISO, Tommy is a strategic leader driving security transformation, risk management, and program maturity. He has led high-performing teams, shaped industry policies, and advised organizations on complex cyber, compliance, and data protection challenges.

Cyber Security Legal Framework in Nepal.pptxGhimire B.R.

Cyber security cyber security cyber security cyber security cyber security cy...pranavbodhak

Gihbli AI and Geo sitution |use/misuse of Ai Technologyzainkhurram1111

End-to-end Assurance for SD-WAN & SASE with ThousandEyesThousandEyes

Let’s Get Slack Certified! 🚀- Slack CommunitySanjeetMishra29

Microsoft Build 2025 takeaways in one presentationDigitalmara

Droidal: AI Agents Revolutionizing HealthcareDroidal LLC

Nix(OS) for Python Developers - PyCon 25 (Bologna, Italia)Peter Bittner

Improving Developer Productivity With DORA, SPACE, and DevExJustin Reock

Protecting Your Sensitive Data with Microsoft Purview - IRMS 2025Nikki Chapple

TrustArc Webinar: Mastering Privacy ContractingTrustArc

Cognitive Chasms - A Typology of GenAI Failure Failure ModesDr. Tathagat Varma

Cybersecurity Fundamentals: Apprentice - Palo Alto CertificateVICTOR MAESTRE RAMIREZ

STKI Israel Market Study 2025 final v1 versionDr. Jimmy Schwarzkopf

Jira Administration Training – Day 1 : IntroductionRavi Teja

Dr Jimmy Schwarzkopf presentation on the SUMMIT 2025 ADr. Jimmy Schwarzkopf

Offshore IT Support: Balancing In-House and Offshore Help Desk Techniciansjohn823664

Measuring Microsoft 365 Copilot and Gen AI SuccessNikki Chapple

Kubernetes Cloud Native Indonesia Meetup - May 2025Prasta Maha

GDG Cloud Southlake #43: Tommy Todd: The Quantum Apocalypse: A Looming Threat...James Anderson

MOV の機械学習システムを支える MLOps 実践【DeNA TechCon 2020 ライブ配信】

1. !1

2. DeNA TechCon 2020 #denatechcon

3. !3

4. !4

5. DeNA TechCon 2020 #denatechcon ✔ ✔ ✔ 5

6. 6

7. 7

8. 8

9. 9 ● ● ● ●

10. 10 々

11. 11

12. 12

13. ● ● !13 ● ●

14. ● ● !14 ● / ● CI/CD ● ● ● ● 々

15. DeNA TechCon 2020 #denatechcon

16. !16 BigQuery 々

17. !17 BigQuery

18. DeNA TechCon 2020 #denatechcon

19. DeNA TechCon 2020 #denatechcon ※

20. DeNA TechCon 2020 #denatechcon ※

21. DeNA TechCon 2020 #denatechcon

22. DeNA TechCon 2020 #denatechcon ※ ？

23. DeNA TechCon 2020 #denatechcon ？

24. DeNA TechCon 2020 #denatechcon 々々 → 々

25. DeNA TechCon 2020 #denatechcon ※

26. !26 ※

27. !27 - ⇒ - - -

28. DeNA TechCon 2020 #denatechcon

29. !29 CI/CD   々

30. !30 CI/CD   々々

31. DeNA TechCon 2020 #denatechcon 々

32. DeNA TechCon 2020 #denatechcon 々 ※

33. !33

34. 々 - - - !34

35. DeNA TechCon 2020 #denatechcon

36. ● ※ ⇒   ※ ●   → ※ !36

37. DeNA TechCon 2020 #denatechcon

38. DeNA TechCon 2020 #denatechcon

39. DeNA TechCon 2020 #denatechcon 々※

40. DeNA TechCon 2020 #denatechcon

42. DeNA TechCon 2020 #denatechcon 々

43. DeNA TechCon 2020 #denatechcon

44. DeNA TechCon 2020 #denatechcon 々 AI Platform TrainingBigQuery

45. DeNA TechCon 2020 #denatechcon 々

46. DeNA TechCon 2020 #denatechcon

47. DeNA TechCon 2020 #denatechcon

48. DeNA TechCon 2020 #denatechcon

49. DeNA TechCon 2020 #denatechcon 々々々々々々

50. DeNA TechCon 2020 #denatechcon 々々々々々々？

51. DeNA TechCon 2020 #denatechcon 々々々々々々

52. DeNA TechCon 2020 #denatechcon ※ 々々

53. DeNA TechCon 2020 #denatechcon ※ 々

54. DeNA TechCon 2020 #denatechcon def task_fail_slack_alert(context): slack_msg = “Task Failed.” failed_alert = SlackWebhookOperator( task_id="xxx", http_conn_id="slack", webhook_token=Variable.get("webhook.token"), message=slack_msg, username="airflow", ) return failed_alert.execute(context=context)

55. DeNA TechCon 2020 #denatechcon <!subteam^ABCDEF123|@user_group>

56. DeNA TechCon 2020 #denatechcon 々々々々々々

57. DeNA TechCon 2020 #denatechcon 々

58. DeNA TechCon 2020 #denatechcon ※

59. DeNA TechCon 2020 #denatechcon

60. DeNA TechCon 2020 #denatechcon

61. DeNA TechCon 2020 #denatechcon

62. DeNA TechCon 2020 #denatechcon 々々

63. DeNA TechCon 2020 #denatechcon 々

64. DeNA TechCon 2020 #denatechcon

65. DeNA TechCon 2020 #denatechcon 々

66. DeNA TechCon 2020 #denatechcon ※

67. DeNA TechCon 2020 #denatechcon → 々

68. DeNA TechCon 2020 #denatechcon 々

69. DeNA TechCon 2020 #denatechcon ※ @pytest.fixture(scope="session", autouse=True) def setup_teardown(request, test_config): # setup # suffix = uuid.uuid4() create_dataset(suffix) yield # teardown # delete_blobs(suffix) delete_dataset(suffix)

70. DeNA TechCon 2020 #denatechcon @pytest.mark.dependency(depends=[“test_preprocess_a”, “test_preprocess_b”]) def test_trainer(test_config): trainer = Trainer( test_config[“xxx”].format(test_config[“hash"]), test_config[“yyy”].format(test_config[“hash"]), ) trainer.train() trainer.upload_model()

71. DeNA TechCon 2020 #denatechcon →

72. DeNA TechCon 2020 #denatechcon 々 →

73. ● ＝ ● ＝ ● !73

74. DeNA TechCon 2020

MOV の機械学習システムを支える MLOps 実践【DeNA TechCon 2020 ライブ配信】

Recommended

More Related Content

What's hot (20)

Similar to MOV の機械学習システムを支える MLOps 実践【DeNA TechCon 2020 ライブ配信】 (20)

More from DeNA (20)

Recently uploaded (20)

MOV の機械学習システムを支える MLOps 実践【DeNA TechCon 2020 ライブ配信】