DEV Community: Mia Chang

AWS NLP Newsletter Archive Directory

Mia Chang — Tue, 29 Nov 2022 14:58:49 +0000

AWS NLP newsletter is curated by AWS NLP Community. We update all you need to know about Natural Language Processing on AWS of the month. Which includes service launch, customer case study, and community event info.

Here comes the archive:

2022

November 2022
October 2022 - NLP Summit Recap
September 2022
August 2022
July 2022 - Summer break. No newsletter released.
June 2022
May 2022
April 2022
March 2022
February 2022
January 2022 - No newsletter released.

2021

December 2021
November 2021
October 2021
September 2021
August 2021 - The First AWS NLP Newsletter

Stay in touch with NLP on AWS

Our contact: aws-nlp@amazon.com
Email us about (1) your awesome project about NLP on AWS, (2) let us know which post in the newsletter helped your NLP journey, (3) other things that you want us to post on the newsletter. Talk to you soon.

AWS NLP Newsletter November 2022

Mia Chang — Mon, 28 Nov 2022 18:18:38 +0000

Hello world. This is the November 2022 edition of the AWS Natural Language Processing (NLP) newsletter covering everything related to NLP at AWS. Feel free to leave comments & share it on your social network.

AWS re:Invent 2022 starts today!

*Explore the AWS AI/ML Services Attendee Guide *– Your complete resource for the AI & machine learning sessions at re:Invent 2022 and share them with your customers.

CheckoutYour guide to AI/ML at AWS re:Invent 2022 ** ** to get a sense of how the AI/ML track is organized and some key sessions.

NLP@AWS Customer Success Stories

ByteDance saves up to 60% on inference costs while reducing latency and increasing throughput using AWS Inferentia"The ByteDance AML team focuses on the research and implementation of cutting-edge ML systems and the heterogenous computing resources they require. We create large-scale training and inference systems for a wide variety of recommender, natural language processing (NLP), and computer vision (CV) models. These models are highly complex and process a huge amount of data from the many content platforms ByteDance operates. Deploying these models requires significant GPU resources, whether in the cloud or on premises. Therefore, the compute costs for these inference systems are quite high […]

Ultimately, after evaluating several options, we chose EC2 Inf1 instances for their better performance/price ratio compared to G4dn instances and NVIDIA T4 on premises. We engaged in a cycle of continuous iteration with the AWS team to unlock the price and performance benefits of Inf1.

Real estate brokerage firm John L. Scott uses Amazon Textract and Amazon Comprehend to strike racially restrictive language from property deeds for homeownersWhen company operating officer Phil McBride joined the company in 2007, one of his initial challenges was to shift the company's public website from an on-premises environment to a cloud-hosted one. According to McBride, a world of resources opened up to John L. Scott once the company started working with AWS to build an easily controlled, cloud-enabled environment.

AI Language Services

2/ AI Language Services

Amazon Textract launches the ability to detect signatures on any document . Amazon Textract is a machine learning service that automatically extracts printed text, handwriting, and data from any document or image. Textract now provides you the capability to detect handwritten signatures, e-signatures, and initials on documents such as loan application forms, checks, claim forms and more. AnalyzeDocument Signatures reduces the need for human reviewers and helps customers reduce costs, save time, and build scalable solutions for document processing.

Introducing DTMF slot settings within Amazon Lex.Amazon Lex is a service for building conversational interfaces into any application using voice and text. With Amazon Lex, you can quickly and easily build conversational bots ("chatbots"), virtual agents, and interactive voice response (IVR) systems. Amazon Lex is excited to launch DTMF-only slot settings and configurable session attributes within the Lex console.

Amazon Translate Enables Tagging Support for Parallel Data and Custom Terminology. Amazon Translate is a neural machine translation service that delivers fast, high-quality, affordable, and customizable language translation. Today, we are launching support of tagging for custom terminology and parallel data resources and then allow/restrict access on them based on the tags.

Amazon Transcribe now supports Thai and Hindi languages for streaming audio. Amazon Transcribe is an automatic speech recognition (ASR) service that makes it easy for you to add speech-to-text capabilities to your applications. Today, we are excited to announce Thai and Hindi language support for streaming audio transcriptions. These new languages expand the coverage of Amazon Transcribe streaming and enable customers to reach a broader global audience.

Amazon Kendra is now FedRAMP High Compliant. Amazon Kendra is now authorized as FedRAMP High in AWS GovCloud (US-West) Region. Amazon Kendra is a highly accurate intelligent search service powered by machine learning. Kendra reimagines enterprise search for your websites and applications so your employees and customers can easily find the content they are looking for, even when it's scattered across multiple locations and content repositories within your organization.

Virtual Private Cloud (VPC) support is generally available for Amazon Polly. Amazon Polly is a service that turns text into lifelike speech, allowing you to create applications that talk, and build entirely new categories of speech-enabled products. Starting today, you can now use Amazon Polly inside an Amazon Virtual Private Cloud (VPC) , instead of connecting over the internet, which allows you to have better control over your network environment.

Intelligent document processing with AWS AI services in the insurance industry. In this two-part series (Part 1, Part 2), they authors take you through how you can automate and intelligently process documents at scale using AWS AI services for an insurance claims processing use case.

Improve data extraction and document processing with Amazon Textract . In this post, the authors demonstrate how to use Amazon Textract to extract meaningful, actionable data from a wide range of complex multi-format PDF files.

NLP on SageMaker

Train gigantic models with near-linear scaling using sharded data parallelism on Amazon SageMaker . Learn how to train a 30B parameter GPT-2 model on SageMaker with ease with Sharded data parallelism on Amazon SageMaker, a new memory-saving distributed training technique in the SageMaker model parallel (SMP) library. Sharded data parallelism is purpose-built for extreme-scale models and uses Amazon in-house MiCS technology under the hood, a science effort to minimize the communication scale by bringing down expensive communication overhead rooted in parameter gathering and gradient synchronization.

AlexaTM 20B is now available in Amazon SageMaker JumpStart . Amazon's state-of-the-art Alexa Teacher Model with 20 billion parameters (AlexaTM 20B) is now available through Amazon SageMaker JumpStart, SageMaker's machine learning hub. You can use AlexaTM 20B for a wide range of industry use-cases, from summarizing financial reports to question answering for customer service chatbots. It can be applied even when there are only a few available training examples, or even none at all. AlexaTM 20B outperforms a 175 billion GPT-3 model on zero-shot learning tasks such as SuperGLUE and shows state-of-the-art performance for multilingual zero-shot tasks such as XNLI.

Deploy BLOOM-176B and OPT-30B on Amazon SageMaker with large model inference Deep Learning Containers and DeepSpeed ** ** Learn how to use a new SageMaker large model inference Deep Learning Container to deploy two of the most popular large NLP models: BigScience's BLOOM-176B and Meta's OPT-30B from the Hugging Face repository.

Transfer learning for TensorFlow text classification models in Amazon SageMaker by SageMaker now provides a new built-in algorithm for text classification using TensorFlow. This supervised learning algorithm supports transfer learning for many pre-trained models available in TensorFlow hub. It takes a piece of text as input and outputs the probability for each of the class labels. You can fine-tune these pre-trained models using transfer learning even when a large corpus of text isn't available.

NLP @ Community

Transformers and Large Language Models (LLMs): A meeting of minds. Conjecture, Amazon Web Services (AWS), and NLP London co-hosted an informal meetup on 22nd November in London, UK to connect Large Language Model (LLM) researchers, practitioners, start-up builders, investors, and business innovators to ignite new ideas and accelerate innovation.

Stability AI release Stable Diffusion 2.0.Stable Diffusion 2.0is out! Read about the new features and improvements it delivers.

Stay in touch with NLP on AWS

Editors

Main Editor - Anastasia Tzeveleka
Reviewed by - Mia Chang

AWS NLP Newsletter - AWS NLP Summit Recap

Mia Chang — Fri, 14 Oct 2022 08:34:59 +0000

What is AWS NLP Summit

The AWS NLP Summit is a two-day, in-person event at the Amazon London Office that took place Oct 5 and Oct 6 2022. It’s a public facing event organised by the AWS Natural Language Processing (NLP) community within AWS. The event covered approaches that you can use to extract insights from text data. For instances, sentiment analysis for financial news, content analysis for educational assets, driving understanding from documents, ..., and more.

Business decision makers learned more about the common use cases and success stories from AWS customers on the first day. Developers and ML practitioners day two, which was packed with workshops and dive deep sessions, was a great opportunity to learn more about the NLP tech stack on AWS.

As you see this post is about “AWS NLP Summit Recap”. We would like to share with you:

Four highlight sessions of the event
Session catalog + LinkedIn links for all our speakers and volunteers
Way to get a copy of the AWS NLP Summit slides :)

Highlights

Keynotes by Craig Saunders
- The keynote from Craig included the knowledge of the conversational AI of Alexa. It gave the audience a glance of what is behind the Alexa service. Craig also shared how Alexa understands the conversation without a subject within a sentence and how Alexa response to users look up the sport game status, analysis how many games are left, and the possibilities for each team to win the game in the future. The session opened up the audience’s understanding of what Alexa actually does behind the scenes.
Start up panel discussion
- Large Language Model (LLM) is super charging AI/ML as an emerging sector. LLM startups have raised over $5bn funding worldwide in the past few years, growing at over 400% YoY. We invited 5 European LLM startup founders, Conjecture, Magic, LightOn, deepset, and Mystic, to share their experience of running LLM startups; How to get started with LLMs - training/fine-tuning/prompting; and the Ethics and the future of AI.
- Relavant posts on LinkedIn:
Customer session [Generating descriptions from structured data with data-to-text models] by NEWTOMS LLC & Antonio Rodriguez
- In this session co-delivered with NEWTOMS, we presented a use case from a Travel & Hospitality customer generating descriptions on hotel properties from structured data using Natural Language Generation (NLG). We saw a demo using SageMaker with a graph database in Neptune, and explored the implementation and its application to other verticals.
Workshop - Distributed Training and Inference for Large Language Models on Amazon SageMaker
- In this session, Emily shared how to train and deploy distributed state of the art models on Amazon SageMaker. From BERT to GPT to BLOOM and more. One of the skill she shared is leveraging distributed strategies like model and data parallelism to optimize for both throughput and accuracy. The LLM is one of the hottest topic of the event, and the session was full-house, and people joined the session with standing spots.

Agenda and the LinkedIn Profile of the speakers

Here we share the session title and speaker list of AWS NLP Summit. Feel free to follow them/connect them on Linkedin.

Keynotes

K01 - AWS NLP Services
Vasi Philomin, Vice President & General Manager, Machine Learning & AI at Amazon

K02 - Accelerate business value with AWS AI/ML
Satish Lakshmanan, Managing Director: AWS Global Head of AI/ML Sales, Biz Development and Solutions Architecture at Amazon Web Services (AWS)

K03 - NLP in Alexa
Craig Saunders, Director of Machine Learning - Alexa AI. at Amazon

Start-up Panels

Yina Ye, ML BD, Startups & VC at Amazon Web Services (AWS)
Connor Leahy, CEO and founder of Conjecture
Eric Steinberger, CEO & Co-Founder at Magic
Igor Carron, Co-Founder and President of LightOn
Timo Möller, Co-Founder of deepset
Paul Hetherington, Co-Founder and CEO of Mystic

Customer speaking

C01 - NEWTOMS LLC - Generating descriptions from structured data with data-to-text models
Antonio Rodriguez, Senior AI/ML Specialist Solutions Architect at Amazon Web Services (AWS)

C02 - DKB ticket intelligence - The trials and tribulations of using BERT
Dr. Thomas Hack, Data Scientist at DKB

C03 - JPMC - Serving BLOOM 176B Model on Amazon SageMaker
Suman Addanki, head of the ML Engineering for AI Services in JPMorgan AI Platform

C04 - Quantiphi - Accelerate NLP deployments by building reusable components
Salman Taherian, AI/ML Partner Solutions Architect at Amazon Web Services (AWS)

C05 - Inawisdom - Driving understanding from Documents and other unstructured data sources
Ines Jurchevici, Snr Partner Dev Specialist at Amazon Web Services (AWS)

C06 - Magic.dev - Large Language Model Training: Self-Managed ML on EC2
Eric Steinberger, CEO & Co-Founder at Magic

Sessions

S01 - Using AWS AI services for building accessibility features and breaking through language barriers
Luca Guida, Solutions Architect at Amazon Web Services (AWS)

S02 - Sentiment Analysis for Financial News using No Code SageMaker Jumpstart and BERT Models
Rahul Sureka, Enterprise Specialist Solutions Architect at Amazon Web Services (AWS)

S03 - Combating misinformation using Amazon Comprehend
Julia Ang, Associate Solutions Architect at Amazon Web Services (AWS)

S04 - Responsible AI: NLP Bias
Hasan Poonawala, Sr ML Specialist Solutions Architect at Amazon Web Services (AWS)
Bharathi Srinivasan, ProServe Data Scientist at Amazon Web Services (AWS)

S05 - Educational content analysis with AIML workloads on AWS
Sebastian Leks, Sr. Service Creation Solutions Architect at Amazon Web Services (AWS)
Mia Chang, AI/ML Specialist Solutions Architect at Amazon Web Services (AWS)

S06 - Visually build conversation flows with Amazon Lex Visual Conversation Builder
Davide Gallitelli, AI/ML Specialist Solutions Architect at Amazon Web Services (AWS)

S07 - Generating art from text with distributed training on Amazon SageMaker
Corey Barrett, Sr. Data Scientist at Amazon Web Services (AWS)
Emily Webber, Principal AI/ML Specialist Solutions Architect at Amazon Web Services (AWS)

S08 - Deploy large NLP models on SageMaker using DJL DeepSpeed inference
Dhawal Patel, Principal AI/ML Specialist Solutions Architect at Amazon Web Services (AWS)
Qingwei Li, Sr. AI/ML Specialist Solutions Architect at Amazon Web Services (AWS)

S09 - NLPOps - Operationalise and automate your NLP pipeline using AWS
Sokratis Kartakis, Sr AI/ML Specialist Solutions Architect at Amazon Web Services (AWS)

S10 - Responsible AI: NLP Explainability
Hasan Poonawala, Sr ML Specialist Solutions Architect at Amazon Web Services (AWS)

S11 - Build highly accurate NLP solutions faster with the open-source AutoGluon and Amazon SageMaker
Anastasia Tzeveleka, Sr AI/ML Solutions Architect, Public Sector at Amazon Web Services (AWS)

S12 - Deeper Document Understanding with Amazon Textract and Amazon SageMaker
Alex Thewsey, Sr AI/ML Solutions Architect at Amazon Web Services (AWS)

S13 - Look Ma, I shrunk BERT (Knowledge Distillation)
Mariano Kamp, Principal Solutions Architect at Amazon Web Services (AWS)

S14 - Accelerate your NLP workloads with Trainium & Inferentia
Matthew McClean, Sr Manager Solutions Architect at Amazon Web Services (AWS)

S15 - Automated mining of emergent catchphrases for advertising content moderation
Ajay Vohra, Principal Solutions Architect, Prototyping at Amazon Web Services (AWS)
Mayank Gupta, Senior SDE at Amazon Web Services (AWS)

S16 - How to deploy and use HuggingFace in an enterprise environment
Sergey Ermolin, Principal Solutions Architect, AI/ML at Amazon Web Services (AWS)

Workshops

W01 - Build a hotel reservation agent
Jake Wen, Solutions Architect at Amazon Web Services (AWS)

W02 - Ask the Architect
Ioan Catana, AI/ML Specialist Solutions Architect at Amazon Web Services (AWS)
Mia Chang, AI/ML Specialist Solutions Architect at Amazon Web Services (AWS)

W03 - Augment your HuggingFace model with Human-in-the-Loop
Georgios Schinas, Specialist Solutions Architect ML at Amazon Web Services (AWS)

W04 - Distributed Training and Inference for Large Language Models on Amazon SageMaker
Emily Webber, Principal AI/ML Specialist Solutions Architect at Amazon Web Services (AWS)

W05 - Text and Images: Multimodal Learning on SageMaker
Uri Rosenberg, Senior STAM – AI/ML at Amazon Web Services (AWS)

W06 - Increase the Search relevance using Semantic embeddings
Praveen Mohan Prasad, STAM - Big Data (Sp) at Amazon Web Services (AWS)

W07 - Hugging Face Model Explainability for customer sentiment text analytics
Tim Wu, Sr. AI Specialist Solutions Architect at Amazon Web Services (AWS)

Was the event recorded?

No, this was an in-person only event. We focus on the face to face experience in this event.

However, if you did not register the event, and you are interested in to have a copy of our available slides. You can request by this slide request form (https://eventbox.dev/survey/WBTQS3Q)

We will send you the copy of the slides by end of 2022 Oct. Note that not all the sessions are available due to PR reason.

Stay in touch with NLP on AWS

Final Thanks.

Organizer team:
Heiko Hotz, Sr. AI/ML Specialist Solutions Architect at Amazon Web Services (AWS)
Mia Chang, AI/ML Specialist Solutions Architect at Amazon Web Services (AWS)
Andrew Kane, WW Tech Leader and Chief Principal Architect AI Language Services at Amazon Web Services (AWS)
Anastasia Tzeveleka, Sr AI/ML Solutions Architect, Public Sector at Amazon Web Services (AWS)
Anna Gruebler, Sr AI/ML Solutions Architect at Amazon Web Services (AWS)
Jon Reade, Sr AI/ML Solutions Architect at Amazon Web Services (AWS)
Yina Ye, ML BD, Startups & VC at Amazon Web Services (AWS)
Premkumar Rangarajan, Principal Solutions Architect at Amazon Web Services (AWS)

Volunteers from AWS and Amazon:
James Park, Fahad Ahmed, Andrew Ellul, Antonio Rodriguez, Dinesh Mane, Lettie Ndlovu, Max Klass, Srivalsan Mannoor Sudhagar, Cosmin Pascu, Maria Mata, Tyree Glasgow-Alexander, Adam Temple, George Mihaila, Ayman Salama, Mihai Dobri, Hasan Basri AKIRMAK, Burak Gozluklu, Yusraa Djaafer, James Yi, Gabrielle Dompreh, Veeresh Shringari, Naufal Mir, Ahmed Raafat, Jenny Vega, Sam Price, Ajay Vohra, Alex Thewsey, Bharathi Srinivasan, Bilal Zafar, Corey Barrett, Emily Webber, Georgios Schinas, Arnaud Lauer, Jake Wen, Luca Guida, Mariano Kamp, Praveen Mohan Prasad, Rahul Sureka, Sebastian Leks, Sergey Ermolin, Tim WU, Uri Rosenberg, Melanie Li, Subham Kumar, Michelle Gibbs, Elina Lesyk, Roop Bains, Sophia Wilson, Brett Hamilton, Sokratis Kartakis

AWS - NLP newsletter October 2021

Mia Chang — Fri, 29 Oct 2021 16:11:23 +0000

Hello world. This is the monthly Natural Language Processing(NLP) newsletter covering everything related to NLP at AWS. This is our third newsletter on Dev.to. If you missed our earlier episode, here are Ep01 and Ep02. Feel free to leave comments, share it on your social network to celebrate this new launch with us!

Service updates about NLP on AWS

Amazon Lex launches progress updates for fulfillment

You can now configure your Amazon Lex bots to provide periodic updates to users while their requests are processed. Customer support conversations often require execution of business logic that can take some time to complete. For example, updating an itinerary on an airline reservation system may take a couple of minutes during peak hours. Typically, support agents put the call on hold and provide periodic updates (e.g., “We are still processing your request; thank you for your patience”) until the request is fulfilled. Now, you can easily configure your bot to automatically provide such periodic updates in a conversation. With progress updates capability, bot builders can quickly enhance the ability of virtual contact center agents and smart assistants.
New AWS Solution: AWS QnABot, a self-service conversational chatbot built on Amazon Lex

The AWS QnABot has now been released as an official AWS Solution Implementation. The AWS QnABot is an open source, multichannel, multi-language conversational chatbot built on Amazon Lex, that responds to your customer’s questions, answers, and feedback. Without programming, the AWS QnABot solution allows customers to quickly deploy self-service conversational AI on multiple channels including their contact centers, websites, social media channels, SMS text messaging, or Amazon Alexa.
Amazon Transcribe now supports custom language models for streaming transcription

Amazon Transcribe will now support custom language models (CLM) for streaming transcription. Amazon Transcribe is an automatic speech recognition (ASR) service that makes it easy for you to add speech-to-text capabilities to your applications. CLM allows you to leverage pre-existing data to build a custom speech engine tailored for your transcription use case. No prior machine learning experience required. AWS ML Blog, Transcribe Documentation.

Translate and analyze text using SQL functions with Amazon Redshift, Amazon Translate, and Amazon Comprehend

You already know how to use Amazon Redshift to transform data using simple SQL commands and built-in functions. Now you can also use Amazon Redshift to translate, analyze, and redact text fields, thanks to Amazon Translate, Amazon Comprehend, and the power of Amazon Redshift supported AWS Lambda user-defined functions (UDFs).
Amazon Comprehend adds two Trusted Advisor checks

Amazon Comprehend now supports two new AWS Trusted Advisor checks to help customers optimize the cost and security of Amazon Comprehend endpoints. Today, Amazon Comprehend checks are available in the AWS Business Support and AWS Enterprise Support plans. The new checks are:
1. Underutilized endpoints: Checks the throughput configuration of your endpoints and generates a warning when they are not actively used for any real-time inference requests;
2. Endpoint permissions: Checks the KMS key permissions for an endpoint whose underlying model was encrypted using customer managed keys. If the customer managed key has been disabled or the key policy has been changed to alter the granted permissions for Amazon Comprehend for any reason, the endpoint availability might be impacted.
Amazon Textract launches TIFF support and adds asynchronous support for receipts and invoices processing

Amazon Textract now supports Tag Image File Format (TIFF) documents in addition to the PNG, JPEG, and PDF formats. Customers can now process TIFF documents either synchronously or asynchronously using any of the following Amazon Textract APIs - DetectDocumentText, StartDocumentAnalysis, StartDocumentTextDetection, AnalyzeDocument, and AnalyzeExpense. Amazon Textract is a machine learning service that automatically extracts printed and handwritten text and data from any document.

NLP on SageMaker

Amazon SageMaker JumpStart introduces new multimodal (long-form text, tabular) financial analysis tools

With this new release, you can use the new set of multimodal financial analysis tools within Amazon SageMaker JumpStart. With these new tools, you can enhance your tabular ML workflows with new insights from financial text documents and potentially help save up to weeks of development time. Using the new SageMaker JumpStart Industry SDK, you can easily retrieve common public financial documents, including SEC filings, and further process financial text documents with features such as summarization and scoring for sentiment, litigiousness, risk, readability etc. In addition, you can access pre-trained language models trained on financial text for transfer learning, and use example notebooks for data retrieval, text feature engineering, multimodal classification and regression models. AWS ML Blog #1, AWS ML Blog #2, AWS ML Blog #3, JumpStart Documentation
Organize product data to your taxonomy with Amazon SageMaker

When companies deal with data that comes from various sources or the collection of this data has changed over time, the data often becomes difficult to organize. Perhaps you have product category names that are similar but don’t match, and on your website you want to surface these products as a group. Therefore, you need to go through the tedious work of manually creating a map from source to target to be able to transform the data into your own taxonomy. In these cases, we’re not talking about a few hundred rows of data, but more often many hundreds of thousands of rows, with new data flowing in regularly. In this post, we discuss how to organize product data to your classification needs with Amazon SageMaker.
Bring structure to diverse documents with Amazon Textract and transformer-based models on Amazon SageMaker

From application forms, to identity documents, recent utility bills, and bank statements, many business processes today still rely on exchanging and analyzing human-readable documents—particularly in industries like financial services and law. In this post, we show how you can use Amazon SageMaker, an end-to-end platform for machine learning (ML), to automate especially challenging document analysis tasks with advanced ML models.

AWS Blog posts, papers, and more

Create a dashboard with SEC text for financial NLP in Amazon SageMaker JumpStart

In this post, the author showed how to curate a dataset of Securities Exchange Commission, SEC filings, use NLP for feature engineering on the dataset, and present the features in a dashboard.

To get started, you can refer to the example notebook in JumpStart titled Dashboarding SEC Filings. You can also refer to the example notebook in JumpStart titled Create a TabText Dataset of SEC Filings in a Single API Call, which contains more details of SEC forms retrieval, summarization, and NLP scoring.
Amazon Science Publication: Sample selection guided by domain and task for cross-domain targeted sentiment analysis

Building supervised targeted sentiment analysis models for a new target domain requires substantial annotation effort since most datasets for this task are domain-specific. Domain adaptation for this task has two dimensions: the nature of targets and the opinion words used to describe sentiment towards the target. We present a data sampling strategy informed by domain differences across these two dimensions with the goal of selecting a small number of examples, thereby minimizing annotation effort. This obtains performance in the 86-100% range compared to the full supervised model using only ∼4-15% of the full training data.

YouTube demo video "Amazon Transcribe video snacks: Using vocabulary filters"

Amazon Transcribe is a automatic speech recognition service that can be used when you have audio and video that contains speech you want to convert to text. You can mask, remove, or tag words you don't want in your transcription results with vocabulary filtering. For example, you can use vocabulary filtering to prevent the display of offensive or profane terms. In the demo, we will customize Transcribe to mask swear words that we recently encountered in a famous play written by William Shakespeare.
4 ways conversational AI and Amazon Lex help the public sector transform customer engagement

Conversational artificial intelligence (AI) and chatbots can be used to transform the customer experience, enhance engagement, improve services, and help scale more simply. Learn how conversational AI and chatbots help public sector organizations.

Community content

Workshop: Getting started with Amazon Sagemaker Train a Hugging Face Transformers and deploy it
Learn how to use Amazon SageMaker to train a Hugging Face Transformer model and deploy it afterward. Prepare and upload a test dataset to S3, prepare a fine-tuning script to be used with Amazon SageMaker Training jobs, Launch a training job and store the trained model into S3, and Deploy the model after successful training. GitHub Repository
October “HuggingFace Blog” entries:
1. Showcase Your Projects in Spaces using Gradio
2. Hosting your Models and Datasets on Hugging Face Spaces using Streamlit
3. Fine-tuning CLIP with Remote Sensing (Satellite) images and captions
4. The Age of Machine Learning As Code Has Arrived
5. Train a Sentence Embedding Model with 1B Training Pairs
6. Large Language Models: A New Moore’s Law?
7. Course Launch Community Event

Upcoming NLP events

Both community events and AWS events

Going Production: Deploying, Scaling & Monitoring Hugging Face Transformer models | Hugging Face
- Tuesday, November 2nd, 2021
- 5:00 PM to 6:00 PM CEST
Pie & AI Suisse - Trustworthiness of AI models: Improving NLP with Causality | Meetup
- Wednesday, November 3, 2021.
- 6:30 PM to 8:00 PM CEST
NLP inference optimization on Amazon SageMaker in NDR conference
- Tuesday, November 09, 2021
- 11:40 AM to 12:20 PM CEST
Analysing Politeness: Can NLP Tools Help? | Meetup
- Wednesday, November 17, 2021
- 8:00 PM to 9:30 PM CEST

Stay in touch with NLP on AWS

AWS - NLP newsletter August 2021

Mia Chang — Fri, 27 Aug 2021 14:19:28 +0000

Hello world. This is the monthly Natural Language Processing(NLP) newsletter covering everything related to NLP at AWS. This is our first newsletter on Dev.to. Special thanks to Ricardo Sueiras helped us make this happen. Feel free to leave comments, share it on your social network to celebrate this new launch with us!

Service updates about NLP on AWS

Amazon Transcribe Call Analytics
Amazon Transcribe Call Analytics is a new machine learning (ML) powered conversation insights API that enables developers to improve customer experience and agent productivity. This API can analyze call recordings to generate turn-by-turn call transcripts and actionable insights for understanding customer-agent interactions, identifying trending issues, and tracking performance metrics. Launch content: AWS News Blog, What's New Post, Webpage, Documentation, GitHub Demo, LinkedIn.
Amazon Connect now works with Amazon Lex V2.
Amazon Lex allows customers to create intelligent chatbots that turn their Amazon Connect contact flows into natural conversations. Amazon Lex V2 console and API enhancements include: 1) support for multiple languages in a simple bot and the ability to manage them as a single resource through the life cycle (build, test, and deploy), 2) ability for end-users to request a bot to wait (“Can you wait while I get my credit card?”), and interrupt a bot in mid-sentence, 3) simplified bot versioning, and 4) new productivity features such as support for saving partially completed bots, bulk upload of sample utterances, and navigation via a dynamic ‘Conversation flow’ for more flexibility and control in the bot design process. Share the news: What’s New post, Amazon Lex Blog post, Amazon Connect documentation, Amazon Lex documentation

AWS Blog posts, papers, and more

Amazon announces new conversational-modeling challenge - Amazon Science

Challenge entry will remain open until September 21, and research teams from academia, industry, and nonprofit and government sectors are welcome to participate. Amazon has open-sourced the development data, evaluation scripts, and baseline systems for challenge participants and other researchers in the field.

Ensuring that new language-processing models don't backslide - Amazon Science

Regression bugs are in your NLP model! This study showed that in updated models, NFRs are often much higher than the total accuracy gains, from two to eight times as high. This implies that simply aiming for greater accuracy improvements in updated models will not ensure a decrease in regression; i.e., improving accuracy and minimizing regression are related but separate learning targets. This post includes mitigations to the regression bugs.

Announcing model improvements and lower annotation limits for Amazon Comprehend custom entity recognition

The post shares how to make improvements to the F1 score while training models with different dataset sampling configurations, including multi-lingual models. With this updated model, Amazon Comprehend makes it easy to train custom entity recognition models. Limits have been lowered to 100 annotations per entity and 250 documents for training, while offering improved accuracy with your models.

Make your audio and video files searchable using Amazon Transcribe and Amazon Kendra | AWS Machine Learning Blog

The combination of Amazon Transcribe and Amazon Kendra enable a scalable, cost-effective solution to make your media files discoverable. You can use the content of your media files to find accurate answers to your users’ questions, whether they’re from text documents or media files, and consume them in their native format.

Analyze customer churn probability using call transcription and customer profiles with Amazon SageMaker

This post explains an end-to-end solution for creating a customer churn prediction model based on customer profiles and customer-agent call transcriptions. Which included training a PyTorch model with a custom script and creating an endpoint for real-time model hosting. Start from create a public-facing API Gateway that can be securely used in your mobile applications or website, then use Amazon Transcribe for batch or real-time transcription of customer-agent conversations, which you can use for training of your model or real-time inference.

Zillow: Near Real-Time Natural Language Processing (NLP) for Customer Interactions

Helping home shoppers connect to the services they need in time can make the difference if they are successful in securing a property or not. In this episode, we explore how Zillow built a natural language processing solution using Amazon Transcribe and leverage the Elastic Container Service to quickly scale their machine learning engine to match customer requests to agents. We dive into how they deploy models into the environment using GitLab pipelines to simplify the job for data scientists.

Community content

NLP to help pregnant mothers in Kenya
In Kenya, 33% of maternal deaths are caused by delays in seeking care, and 55% of maternal deaths are caused by delays in action or inadequate care by providers. Jacaranda Health is employing NLP and dialogue system techniques to help mothers experience childbirth safely and with respect, and to help newborns get a safe start in life.
HuggingFace Processing Jobs on Amazon SageMaker
Prepare text data for your NLP pipeline in a scalable and reproducible way. This has two principal benefits: (1) For large datasets, data preparation can take a long time. Choosing dedicated EC2 instances allows us to pick the right processing power for the task at hand. (2) Codifying the data preparation via a processing job enables us to integrate the data processing step into a CI/CD pipeline for NLP tasks in a scalable and reproducible way.
How Good Is Your NLP Model Really?
How to evaluate NLP models with Amazon SageMaker Processing jobs for Hugging Face’s Transformer models. NLP model evaluation can be resource-intensive, especially when it comes to Transformer models that benefit greatly from GPU acceleration. In the post we will then see that we can speed up this process up to 267(!) times by using SageMaker’s Hugging Face Processing jobs.

Code samples

HuggingFace

Amazon SageMaker enables customers to train, fine-tune, and run inference using Hugging Face models for Natural Language Processing(NLP) on SageMaker.

NLP with BlazingText with Amazon SageMaker and AWS Lambda

This example illustrates how to use a BlazingText text classification training with SageMaker, and serving with AWS Lambda for both supervised (text classification) and unsupervised (Word2Vec) modes. The repository comes with Jupyter notebook, docker container, and events file you can work with.

GitHub: Train a BlazingText text classification algorithm in SageMaker, inference with AWS Lambda

Upcoming events

Best Practices in Conversational AI Design
Sep 01, 2021 | 07:00 PM CEST

Join conversational design leaders from Amazon, Alexa, and AWS as we discuss best practices in conversational AI design. Building conversational interfaces can be challenging given the free-form nature of communication and unstructured data. Users can say whatever they like, however they like. It is quite a bit different from web and mobile design. Our experts will cover the best practices in conversational design, including how to design for voice assistants versus text chatbots, handling fallbacks gracefully, the role of context and personalization, how to guide the users along the happy path to successful engagement, tips for creating an intuitive flow, and more.