Active learning workflow for Amazon Comprehend custom classification models – Part 1

Amazon Comprehend  Custom Classification API enables you to easily build custom text classification models using your business-specific labels without learning ML. For example, your customer support organization can use Custom Classification to automatically categorize inbound requests by problem type based on how the customer has described the issue.  You can use custom classifiers to automatically label support emails with appropriate issue types, routing customer phone calls to the right agents, and categorizing social media posts into user segments.

For custom classification, you start by creating a training job with a ground truth dataset comprising a collection of text and corresponding category labels. Upon completing the job, you have a classifier that can classify any new text into one or more named categories. When the custom classification model classifies a new unlabeled text document, it predicts what it has learned from the training data. Sometimes you may not have a training dataset with various language patterns, or once you deploy the model, you start seeing completely new data patterns. In these cases, the model may not be able to classify these new data patterns accurately. How can we ensure continuous model training to keep it up to date with new data and patterns?

In this two part blog series, we discuss an architecture pattern that allows you to build an active learning workflow for Amazon Comprehend custom classification models. The first post will describe a workflow comprising real-time classification, feedback pipelines and human review workflows using Amazon Augmented AI (Amazon A2I). The second post will cover the automated model building using the human reviewed data, selecting the best model, and automated deployment of an endpoint of the chosen model.

Feedback loops play a pivotal role in keeping the models up to date. This feedback helps the models learn about their misclassifications and learn the right ones. This process of teaching the models continuously through feedback and deploying them is called active learning.

For every prediction Amazon Comprehend Custom Classification makes, it also gives a confidence score associated with its prediction. This architecture proposes that you set an acceptable threshold and only accept the predictions with a confidence score that exceeds the threshold. All the predictions that have a confidence score less than the desired threshold are flagged for human review. The human decides whether to accept the model’s prediction or correct it.

In some instances, the model may be confident about its predictions, but the classification might be wrong. In these scenarios, the end-user applications that receive the model predictions can request explicit feedback from its users on the prediction quality. A human moderator reviews this explicit feedback and reclassifies instances where the feedback was negative. This process of generating human-verified data and using it for model retraining helps keep the models up to date, reduce data drift, and achieve higher model accuracy.

Feedback Workflow Architecture.

In this section, we discuss an architectural pattern for implementing an end-to-end active learning workflow for custom classification models in Amazon Comprehend using Amazon A2I. The active learning workflow comprises the following components:

1. Real-time classification
2. Feedback loops
3. Human classification
4. Model building
5. Model selection
6. Model deployment

The following diagram illustrates this architecture covering the first three components. In the following sections, we walk you through each step in the workflow.

Real-time classification

To use custom classification in Amazon Comprehend, you need to create a custom classification job that reads a ground truth dataset from an Amazon Simple Storage Service (Amazon S3) bucket and builds a classification model. After the model builds successfully, you can create an endpoint that allows you to make real-time classifications of unlabeled text. This stage is represented by steps 1–3 in the preceding architecture:

1. The end-user application calls an API Gateway endpoint with a text that needs to be classified.
2. The API Gateway endpoint then calls an AWS Lambda function configured to call an Amazon Comprehend endpoint.
3. The Lambda function calls the Amazon Comprehend endpoint, which returns the unlabeled text classification and a confidence score.

Feedback collection

When the endpoint returns the classification and the confidence score during the real-time classification, you can send instances with low-confidence scores to human review. This type of feedback is called implicit feedback.

4. The Lambda function sends the implicit feedback to an Amazon Kinesis Data Firehose.

The other type of feedback is called explicit feedback and comes from the application’s end-users that use the custom classification feature. This type of feedback comprises the instances of text where the user wasn’t happy with the prediction. Explicit feedback can be sent either in real-time through an API or a batch process.

5. End-users of the application submit explicit real-time feedback through an API Gateway endpoint.
6. The Lambda function backing the API endpoint transforms the data into a standard feedback format and writes it to the Kinesis Data Firehose delivery stream.
7. End-users of the application can also submit explicit feedback as a batch file by uploading it to an S3 bucket.
8. A trigger configured on the S3 bucket triggers a Lambda function.
9. The Lambda function transforms the data into a standard feedback format and writes it to the delivery stream.
10. Both the implicit and explicit feedback data gets sent to a delivery stream in a standard format. All this data is buffered and written to an S3 bucket.

Human classification

The human classification stage includes the following steps:

11. A trigger configured on the feedback bucket in Step 10 invokes a Lambda function.
12. The Lambda function creates Amazon A2I human review tasks for all the feedback data received.
13. Workers assigned to the classification jobs log in to the human review portal and either approve the classification by the model or classify the text with the right labels.
14. After the human review, all these instances are stored in an S3 bucket and used for retraining the models. Part 2 of this series covers the retraining workflow.

Solution overview

The next few sections of the post go over how to set up this architecture in your AWS account. We classify news into four categories: World, Sports, Business, and Sci/Tech, using the AG News dataset for custom classification, and set up the implicit and explicit feedback loop. You need to complete two manual steps:

1. Create an Amazon Comprehend custom classifier and an endpoint.
2. Create an Amazon SageMaker private workforce, worker task template, and human review workflow.

After this, you run the provided AWS CloudFormation template to set up the rest of the architecture.

Prerequisites

Before you get started, download the dataset and upload it to Amazon S3. This dataset comprises a collection of news articles and their corresponding category labels. We have created a training dataset called train.csv from the original dataset and made it available for download.

The following screenshot shows a sample of the train.csv file.

After you download the train.csv file, upload it to an S3 bucket in your account for reference during training. For more information about uploading files, see How do I upload files and folders to an S3 bucket?

Creating a custom classifier and an endpoint

To create your classifier for classifying news, complete the following steps:

1. On the Amazon Comprehend console, choose Custom Classification.
2. Choose Train classifier.
3. For Name, enter news-classifier-demo.
4. Select Using Multi-class mode.
5. For Training data S3 location, enter the path for train.csv in your S3 bucket, for example, s3://<your-bucketname>/train.csv.
6. For Output data S3 location, enter the S3 bucket path where you want the output, such as s3://<your-bucketname>/.
7. For IAM role, select Create an IAM role.
8. For Permissions to access, choose Input and output (if specified) S3 bucket.
9. For Name suffix, enter ComprehendCustom.

10. Scroll down and choose Train Classifier to start the training process.

The training takes some time to complete. You can either wait to create an endpoint or come back to this step later after finishing the steps in the section Creating a private workforce, worker task template, and human review workflow.

Creating a custom classifier real-time endpoint

To create your endpoint, complete the following steps:

1. On the Amazon Comprehend console, choose Custom Classification.
2. From the Classifiers list, choose the name of the custom model for which you want to create the endpoint and select your model news-classifier-demo.
3. From the Actions drop-down menu, choose Create endpoint.
4. For Endpoint name, enter classify-news-endpoint and give it one inference unit.
5. Choose Create endpoint
6. Copy the endpoint ARN as shown in the following screenshot. You use it when running the CloudFormation template in a future step.

Creating a private workforce, worker task template, and human review workflow.

This section walks you through creating a private workforce in Amazon SageMaker, a worker task template, and your human review workflow.

Creating a labeling workforce

1. For this post, you will create a private work team and add only one user (you) to it. For instructions, see Create a Private Workforce (Amazon SageMaker Console).
2. Once the user accepts the invitation, you will have to add him to the workforce. For instructions, see the Add a Worker to a Work Team section the Manage a Workforce (Amazon SageMaker Console)

Creating a worker task template

To create a worker task template, complete the following steps:

1. On the Amazon A2I console, choose Worker task templates.
2. Choose to Create a template.
3. For Template name, enter custom-classification-template.
4. For Template type, choose Custom,
5. In the Template editor, enter the following GitHub UI template code.
6. Choose Create.

Creating a human review workflow

To create your human review workflow, complete the following steps:

1. On the Amazon A2I console, choose Human review workflows.
2. Choose Create human review workflow.
3. For Name, enter classify-workflow.
4. Specify an S3 bucket to store output: s3://<your bucketname>/.

Use the same bucket where you downloaded your train.csv in the prerequisite step.

5. For IAM role, select Create a new role.
6. For Task type, choose Custom.
7. Under Worker task template creation, select the custom classification template you created.
8. For Task description, enter Read the instructions and review the document.
9. Under Workers, select Private.
10. Use the drop-down list to choose the private team that you created.
11. Choose Create.
12. Copy the workflow ARN (see the following screenshot) to use when initializing the CloudFormation parameters.

Deploying the CloudFormation template to set up active learning feedback

Now that you have completed the manual steps, you can run the CloudFormation template to set up this architecture’s building blocks, including the real-time classification, feedback collection, and the human classification.

Before deploying the CloudFormation template, make sure you have the following to pass as parameters:

• Custom classifier endpoint ARN
• Amazon A2I workflow ARN

1. Choose Launch Stack:

2. Enter the following parameters:
   1. ComprehendEndpointARN – The endpoint ARN you copied.
   2. HumanReviewWorkflowARN – The workflow ARN you copied.
   3. ComrehendClassificationScoreThreshold – Enter 0.5, which means a 50% threshold for low confidence score.

3. Choose Next until the Capabilities
4. Select the check-box to provide acknowledgment to AWS CloudFormation to create AWS Identity and Access Management (IAM) resources and expand the template.

For more information about these resources, see AWS IAM resources.

5. Choose Create stack.

Wait until the status of the stack changes from CREATE_IN_PROGRESS to CREATE_COMPLETE.

6. On the Outputs tab of the stack (see the following screenshot), copy the value for  BatchUploadS3Bucket, FeedbackAPIGatewayID, and TextClassificationAPIGatewayID to interact with the feedback loop.
7. Both the TextClassificationAPI and FeedbackAPI will require and API key to interact with them. The Cloudformtion output ApiGWKey refers to the name of the API key. Currently this API key is associated with a usage plan that allows 2000 requests per month.
8. On the API Gateway console, choose either the TextClassification API or the the FeedbackAPI. Choose API Keys from the left navigation. Choose your API key from step 7. Expand the API key section in the right pane and copy the value.

9. You can manage the usage plan by following the instructions on, Create, configure, and test usage plans with the API Gateway console.
10. You can also add fine grained authentication and authorization to your APIs. For more information on securing your APIs, you can follow instructions on Controlling and managing access to a REST API in API Gateway.

Testing the feedback loop

In this section, we walk you through testing your feedback loop, including real-time classification, implicit and explicit feedback, and human review tasks.

Real-time classification

To interact and test these APIs, you need to download Postman.

The API Gateway endpoint receives an unlabeled text document from a client application and internally calls the custom classification endpoint, which returns the predicted label and a confidence score.

1. Open Postman and enter the TextClassificationAPIGateway URL in POST method.
2. In the Headers section, configure the API key.  x-api-key :  << Your API key >>
3. In the text field, enter the following JSON code (make sure you have JSON selected and enable raw):

{"classifier":"<your custom classifier name>", "sentence":"MS Dhoni retires and a billion people had mixed feelings."}

3. Choose Send.

You get a response back with a confidence score and class, as seen in the following screenshot.

Implicit feedback

When the endpoint returns the classification and the confidence score during the real-time classification, you can route all the instances where the confidence score doesn’t meet the threshold to human review. This type of feedback is called implicit feedback. For this post, we set the threshold as 0.5 as an input to the CloudFormation stack parameter.

You can change this threshold when deploying the CloudFormation template based on your needs.

Explicit feedback

The explicit feedback comes from the end-users of the application that uses the custom classification feature. This type of feedback comprises the instances of text where the user wasn’t happy with the prediction. You can send the predicted label by the model’s explicit feedback through the following methods:

• Real time through an API, which is usually triggered through a like/dislike button on a UI.
• Batch process, where a file with a collection of misclassified utterances is put together based on a user survey conducted by the customer outreach team.

Invoking the explicit real-time feedback loop

To test the Feedback API, complete the following steps:

1. Open Postman and enter the FeedbackAPIGatewayID value from your CloudFormation stack output in POST method.
2. In the Headers section, configure the API key.  x-api-key :  << Your API key >>
3. In the text field, enter the following JSON code (for classifier, enter the classifier you created, such as news-classifier-demo, and make sure you have JSON selected and enable raw):

{"classifier":"<your custom classifier name>","sentence":"Sachin is Indian Cricketer."}

3. Choose Send.

Submitting explicit feedback as a batch file

Download the following test feedback JSON file, populate it with your data, and upload it into the BatchUploadS3Bucket created when you deployed your CloudFormation template. The following code shows some sample data in the file:

{
   "classifier":"news-classifier-demo",
   "sentences":[
      "US music firms take legal action against 754 computer users alleged to illegally swap music online.",
      "A gamer spends $26,500 on a virtual island that exists only in a PC role-playing game."
   ]
}

Uploading the file triggers the Lambda function that starts your human review loop.

Human review tasks

All the feedback collected through the implicit and explicit methods is sent for human classification. The labeling workforce can include Amazon Mechanical Turk, private teams, or AWS Marketplace vendors. For this post, we create a private workforce. The URL to the labeling portal is located on the Amazon SageMaker console, on the Labeling workforces page, on the Private tab.

After you log in, you can see the human review tasks assigned to you. Select the task to complete and choose Start working.

You see the tasks displayed based on the worker template used when creating the human workflow.

After you complete the human classification and submit the tasks, the human-reviewed data is stored in the S3 bucket you configured when creating the human review workflow. Go to Amazon Sagemaker-> Human review workflows->output location:

This human-reviewed data is used to retrain the custom classification model to learn newer patterns and improve its overall accuracy. Below is screenshot of the human annotated output file output.json in S3 bucket:

The process of retraining the models with human-reviewed data, selecting the best model, and automatically deploying the new endpoints completes the active learning workflow. We cover these remaining steps in Part 2 of this series.

Cleaning up

To remove all resources created throughout this process and prevent additional costs, complete the following steps:

1. On the Amazon S3 console, delete the S3 bucket that contains the training dataset.
2. On the Amazon Comprehend console, delete the endpoint and the classifier.
3. On the Amazon A2I console, delete the human review workflow, worker template, and the private workforce.
4. On the AWS CloudFormation console, delete the stack you created. (This removes the resources the CloudFormation template created.)

Conclusion

Amazon Comprehend helps you build scalable and accurate natural language processing capabilities without any machine learning experience. This post provides a reusable pattern and infrastructure for active learning workflows for custom classification models. The feedback pipelines and human review workflow help the custom classifier learn new data patterns continuously. The second part of this series covers the automatic model building, selection, and deployment of custom classification models.

For more information, see Custom Classification. You can discover other Amazon Comprehend features and get inspiration from other AWS blog posts about how to use Amazon Comprehend beyond classification.

---

About the Authors

 Shanthan Kesharaju is a Senior Architect in the AWS ProServe team. He helps our customers with AI/ML strategy, architecture, and develop products with a purpose. Shanthan has an MBA in Marketing from Duke University and an MS in Management Information Systems from Oklahoma State University.

 

 

 

Mona Mona is an AI/ML Specialist Solutions Architect based out of Arlington, VA. She works with World Wide Public Sector team and helps customers adopt machine learning on a large scale. She is passionate about NLP and ML Explainability areas in AI/ML.

 

 

 

 

Joyson Neville Lewis obtained his master’s in Information Technology from Rutgers University in 2018. He has worked as a Software/Data engineer before diving into the Conversational AI domain in 2019, where he works with companies to connect the dots between business and AI using voice and chatbot solutions. Joyson joined Amazon Web Services in February of 2018 as a Big Data Consultant for AWS Professional Services team in NYC.

from AWS Machine Learning Blog https://ift.tt/30aRkwB
via A.I .Kung Fu

California County Enlists Social Media to Thwart a Misleading Election Photo

Officials moved swiftly when old images of discarded mail-in ballots were circulated and portrayed as new pictures.

from NYT > Technology https://ift.tt/331IKSI
via A.I .Kung Fu

Future AR wearables from Facebook, Apple, and Amazon will track how we move and react in real time, setting up a privacy minefield (Karissa Bell/Engadget)

Karissa Bell / Engadget:
Future AR wearables from Facebook, Apple, and Amazon will track how we move and react in real time, setting up a privacy minefield  —  Facebook recently gave us our best glimpse yet into its augmented reality plans.  The company will be piloting a new set of glasses that will lay the groundwork for an eventual consumer-ready product.

from Techmeme https://ift.tt/3j37z6w
via A.I .Kung Fu

How worried should we be about deadly cyber-attacks?

Experts have been warning for years that it's not a matter of if, but when, hackers will kill somebody.

from BBC News - Technology https://ift.tt/3hXDTq5
via A.I .Kung Fu

How a ballerina became a gamer and entrepreneur: The Mari Takahashi origin story - CNET

On CNET's I'm So Obsessed podcast, Takahashi explains her unusual career path, being on Survivor, working for the YouTube giant Smosh and becoming, and now creating, her own content.

from CNET News https://ift.tt/3j2OlxF
via A.I .Kung Fu

Judge gives Trump administration until Friday to defend or delay TikTok ban - CNET

A ban against the Chinese-owned short form video app is set to take effect Sunday.

from CNET News https://ift.tt/2RTTbS9
via A.I .Kung Fu

Without Evidence, Right-Wing Commentators Link Soros to Louisville U-Haul

A spokeswoman for the liberal financier said he had “absolutely not” paid for the rental of a truck used in a protest in Louisville.

from NYT > Technology https://ift.tt/3ctsrRP
via A.I .Kung Fu

Google Services Go Down in Some Parts of U.S.

People experienced outages of services like Gmail, YouTube and Google Meet.

from NYT > Technology https://ift.tt/3cw2Esf
via A.I .Kung Fu

Chinese startup Showmac Tech, which makes eSIMs for Xiaomi IoT devices, raises $15M Series A+ led by Addor Capital (Rita Liao/TechCrunch)

Rita Liao / TechCrunch:
Chinese startup Showmac Tech, which makes eSIMs for Xiaomi IoT devices, raises $15M Series A+ led by Addor Capital  —  Connectivity is vital to a future managed and shaped by smart hardware, and Chinese startup Showmac Tech is proposing eSIMs as the infrastructure solution for seamless …

from Techmeme https://ift.tt/3mU4DLU
via A.I .Kung Fu

Amazon Luna: How to get early access to the cloud gaming platform - CNET

Luna doesn't have a release date, but you can register for early access right now.

from CNET News https://ift.tt/3j2RBcm
via A.I .Kung Fu

Coronavirus lockdowns made Alexa vital, and Amazon knows it - CNET

Amazon introduces a bunch of new gadgets and services to help with voice conferencing, home security and video streaming.

from CNET News https://ift.tt/330Dpv9
via A.I .Kung Fu

The Boys TV series gets a new spinoff series at Amazon - CNET

The new series, to be set in a superhero college, is partially inspired by The Hunger Games.

from CNET News https://ift.tt/3048egA
via A.I .Kung Fu

Facebook Takes Down Networks Linked to Russian Disinformation

The social network said it was moving proactively to dismantle infrastructure Russia could use against the American presidential election.

from NYT > Technology https://ift.tt/332r21E
via A.I .Kung Fu

Sources: ahead of public market debut on Sept. 30, Palantir has told investors that its shares could start trading at $10 apiece, valuing the company ~$22B (Wall Street Journal)

Wall Street Journal:
Sources: ahead of public market debut on Sept. 30, Palantir has told investors that its shares could start trading at $10 apiece, valuing the company ~$22B  —  Bankers have told investors stock could start trading at around $10, sources say  —  Palantir Technologies Inc. is expected to fetch …

from Techmeme https://ift.tt/3mPs1d8
via A.I .Kung Fu

Former director at Facebook, Tim Kendall, told Congress Facebook "took a page from Big Tobacco's playbook, working to make our offering addictive at the outset" (Kate Cox/Ars Technica)

Kate Cox / Ars Technica:
Former director at Facebook, Tim Kendall, told Congress Facebook “took a page from Big Tobacco's playbook, working to make our offering addictive at the outset”  —  “At worst, I fear we are pushing ourselves to the brink of a civil war,” he added.

from Techmeme https://ift.tt/3i4I7Mq
via A.I .Kung Fu

Clearview AI SEC filing revealed $8.6M in equity sales from undisclosed investors and identified two previously unknown board members (BuzzFeed News)

BuzzFeed News:
Clearview AI SEC filing revealed $8.6M in equity sales from undisclosed investors and identified two previously unknown board members  —  Controversial facial recognition company Clearview AI — which has built a database of more than 3 billion images taken from Facebook, Instagram …

from Techmeme https://ift.tt/3iWIMRo
via A.I .Kung Fu

Amazon says it monitors internal listservs to gather feedback at scale, after an AWS employee claimed the listservs are being monitored for labor organizing (Lauren Kaori Gurley/VICE)

Lauren Kaori Gurley / VICE:
Amazon says it monitors internal listservs to gather feedback at scale, after an AWS employee claimed the listservs are being monitored for labor organizing  —  An Amazon spokesperson confirmed that the company monitors dozens of internal listservs but said that the program did not exist for the purposes that the employee described.

from Techmeme https://ift.tt/3mQpdNa
via A.I .Kung Fu

Amazon employees fear HR is targeting minority and activism groups in email monitoring program

Amazon employees participating in the Global Climate Strike in Seattle, Washington in September 2020 | Photo by Karen Ducey/Getty Images

Some workers are raising concerns in light of larger tensions over labor organizing.

Some Amazon employees are furious after they discovered the company’s HR department appears to be quietly monitoring a subset of listservs dedicated to employees who are minorities and those who are involved in activism.

Earlier this week, a group of Amazon employees discovered that an email alias affiliated with Amazon’s HR team had subscribed to 78 listservs at Amazon, the majority of which are related to underrepresented employees and employee activism issues, such as climate change, Black employee networking, and Muslim employees. Amazon has thousands of internal emailing lists where employees discuss common interests and projects, so the dozens of listservs that the alias was subscribed to are a small subset of the total groups that exist.

Amazon denies that its HR teams were tracking emails to monitor organizing, and told Recode that it subscribed to the groups to monitor employee feedback on company culture.

Some of the listservs that Amazon HR appears to be monitoring have been used for employee organizing around controversial issues at the company in recent months, such as Amazon warehouse worker rights, corporate carbon emissions, and military technology. Others are less political, such as groups for women in engineering and employees who are parents.

Recode spoke with six Amazon employees who said they’re alarmed by the email monitoring when they consider how Amazon has been increasingly cracking down on worker organizing. These employees spoke on the condition of anonymity because of Amazon’s policies against speaking to the press without management’s approval.

They told Recode that dozens of other colleagues have also posted messages on Amazon’s internal forums expressing concern about the monitoring, and many more are likely also upset but too scared to speak out publicly.

“Most people would just read and go quiet,” said one employee. “It doesn’t seem smart to engage when we’re being told that we are being tracked.”

Just last month, Bloomberg News reported that Amazon had posted a job listing (which it later removed) for analysts to research “labor organizing threats against the company.”

“If this is what it looks like ... then this is a specific targeting of non-white male groups as potential threats to be observed,” said one employee. “It means that the people responsible for that at Amazon believe that women and people of color are suspicious and threats to the company.”

After discovering the alias, an Amazon employee sent a message to all 78 listservs informing them that the company seemed to be watching their activities. Vice first reported on an Amazon employee warning their colleagues about the monitoring.

“Good day! If you are a moderator or a user of this list, please note that it is being explicitly watched,” started the email. It goes on to state, “Without editorialising (sic), it is difficult to read this project and its initial posting date without also considering the context of the recent job posting for which Amazon has come under fire.” (That’s a reference to the “labor organizing threats” analyst position.)

The email noted that while the Muslim employee listserv was on the list of groups that HR was watching, the Christian one was not.

Amazon spokesperson Jaci Anderson said the practice was intended to gather employee feedback to help improve company practices, and that Amazon does not link feedback from the emails to individual employees. She added that the company chooses which listservs to monitor based on the size and level of activity of the employee group, and for no other reason.

“We continually work to improve the Amazon employee experience, and with hundreds of thousands of employees located around the world, we use several methods to gather feedback at scale,” Amazon spokesperson Jaci Anderson said in a statement. “The anonymized feedback that is sometimes shared from these open email forums has helped us improve our employee benefits, further strengthen our COVID-19 procedures, and improve the overall Amazon employee experience.”

One of the Amazon employees who appears to be linked to the email alias is a data analyst in the employee relations division of HR. And because some employee relations roles at Amazon involve mitigating the risk of unionization in its massive warehouse network, this detail has fueled corporate employees’ concerns about the listserv monitoring.

According to documents reviewed by Recode, the email account subscribed to these groups also appears to be linked to a larger data visualization project run by Amazon’s employee relations team called “SPOC” (geoSPatial Operating Console) which involves monitoring threats to Amazon’s operations — including unionization. Anderson, the Amazon spokesperson, said the program monitors all types of external activity that impacts the safety and well-being of its employees, from weather events to power outages, and is not intended to favor or target one type of external threat over another.

Shortly after an Amazon employee emailed documents about the broader SPOC monitoring project to dozens of employee listservs, those documents were deleted from Amazon’s internal network that’s broadly available to employees.

In April, Business Insider reported that Amazon was tracking Whole Foods’ workers unionization efforts in a geographic heat map. And Recode has previously reported how as far back as the early 2000s, Amazon has previously tracked worker organizing at its warehouses before using excel to make heat maps.

“It’s disappointing but not surprising,” said one Amazon engineering manager about HR tracking listserv emails and broader anti-unionization monitoring efforts. “We are working in corporate America at one of the largest and most technically advanced companies in the world. Always assume big brother is watching.”

This employee said they wish the company was as employee-obsessed as it is customer-obsessed, but also acknowledged that the company “overall takes care of their [corporate] employees very well.”

Amazon has faced growing tensions within both its corporate and blue-collar workforces in recent years. Tensions peaked during the Covid-19 pandemic as many of its warehouse workers complained about working conditions and pay — and some started nascent talks about unionization. The company came under serious scrutiny when it fired Christian Smalls, a former Staten Island warehouse worker who was organizing his colleagues for safer working conditions at the beginning of the pandemic, and after a report emerged that a top Amazon lawyer called Smalls, who is Black, “not smart, or articulate” in an executive meeting. Amazon said it fired Smalls for violating social distancing rules.

The Smalls case, among other factors, prompted many of the companies’ tech employees to advocate for greater workplace protections for Amazon warehouse and delivery workers. The company responded by providing some incremental benefits, such as temporarily raising pay for fulfillment center employees and offering more time off. At the same time, Amazon also fired several corporate employees leading internal activism efforts, who had organized their tech colleagues in solidarity with Smalls and other warehouse workers.

“After firing a number of employees organizing for better COVID-19 protections, and getting caught calling a Black organizer ‘not articulate,’ it seems like Bezos decided it was time to form an internal anti-worker police force to frighten employees into shutting up,” one Amazon software engineer told Recode. “That open feeling I had as a tech worker able to freely discuss ideas with coworkers has vanished so fast it’s made my head spin. There’s a real culture shift happening, and it sucks.”

In the months since Smalls’ firing, Amazon employee activism has quieted down, at least publicly. But workers’ fears about the company targeting and monitoring minority and activist employees show that tensions still run deep at the company, and if anything, its workers’ trust in their employer continues to erode.

from Vox - Recode https://ift.tt/3mQU6AY
via A.I .Kung Fu

The first black hole ever photographed is apparently 'wobbling' - CNET

This supermassive black hole likes to wiggle.

from CNET News https://ift.tt/3056rrB
via A.I .Kung Fu

UFC 253: Israel Adesanya vs. Paulo Costa -- how to watch online, start time and full fight card - CNET

How to watch two undefeated superstars of MMA fight for the UFC middleweight title at UFC 253.

from CNET News https://ift.tt/32YO3SU
via A.I .Kung Fu

This might be the safest way to see a movie in a theater these days - CNET

Alamo Drafthouse is allowing moviegoers to rent out entire theaters for trusted friends and family.

from CNET News https://ift.tt/364kdOQ
via A.I .Kung Fu

The Ring Always Home Cam Flies Around Inside Your House

Look at this freaking drone.

from Wired https://ift.tt/2RYslbi
via A.I .Kung Fu

'One day everyone will use China's digital currency'

China plans a digital version of its currency, which some say could become a big global payment system.

from BBC News - Technology https://ift.tt/305Twpi
via A.I .Kung Fu

The QX60 Monograph concept is a thinly veiled look at Infiniti's next SUV - Roadshow

This technically isn't the new Infiniti QX60 crossover, but it's really, really close.

from CNET News https://ift.tt/3kYsisP
via A.I .Kung Fu

Apple confirms it has acquired cross-platform podcast app Scout FM, which created podcast stations for users based on listening history, earlier this year (Mark Gurman/Bloomberg)

Mark Gurman / Bloomberg:
Apple confirms it has acquired cross-platform podcast app Scout FM, which created podcast stations for users based on listening history, earlier this year  —  - Company acquired Scout FM earlier this year in podcasting push  — Acquisition continues rapid pace of Apple deals this year

from Techmeme https://ift.tt/3csbkjc
via A.I .Kung Fu

Fall Guys are apparently horrific, nightmarish monsters on the inside - CNET

Their insides look much creepier than their adorable outside exterior.

from CNET News https://ift.tt/2G3Nbna
via A.I .Kung Fu

PlayStation 5 launch games: Every PS5 title you can play on day one - CNET

Spider-Man, Demon's Souls and more.

from CNET News https://ift.tt/33TTWzZ
via A.I .Kung Fu

NHS COVID-19, a contact tracing app created by the UK government, launches in England and Wales with Exposure Notification API from Apple and Google (Juli Clover/MacRumors)

Juli Clover / MacRumors:
NHS COVID-19, a contact tracing app created by the UK government, launches in England and Wales with Exposure Notification API from Apple and Google  —  NHS COVID-19, a contact tracing app created by the UK government, is rolling out to residents of England and Wales as of today, with the app available from the iOS App Store.

from Techmeme https://ift.tt/2G4r5kt
via A.I .Kung Fu

WIRED25 Day 2: How to Build a More Resilient World

Guests like Nextdoor CEO Sarah Friar, hacker Matt Mitchell, and journalist Maria Ressa talk about shifting power dynamics—and changing who's in control.

from Wired https://ift.tt/32XhqF9
via A.I .Kung Fu

GOAT Group, operator of online sneaker marketplace GOAT, has raised $100M Series E from D1 Capital Partners, source says at a $1.75B valuation (Sahil Patel/Wall Street Journal)

Sahil Patel / Wall Street Journal:
GOAT Group, operator of online sneaker marketplace GOAT, has raised $100M Series E from D1 Capital Partners, source says at a $1.75B valuation  —  New funding round values the company at $1.75 billion, according to a person familiar with the deal  —  Goat Group, best known for its online marketplace …

from Techmeme https://ift.tt/33UL90L
via A.I .Kung Fu

A rare blue moon will bring a Halloween 2020 treat to the skies - CNET

For the first time since World War II, people in all parts of the world will be able to see the Oct. 31 display.

from CNET News https://ift.tt/33Xnm0j
via A.I .Kung Fu

Amazon launches "Climate Pledge Friendly" label in search results for more than 25K products that have one or more of 19 different sustainability certifications (Nick Statt/The Verge)

Nick Statt / The Verge:
Amazon launches “Climate Pledge Friendly” label in search results for more than 25K products that have one or more of 19 different sustainability certifications  —  The company is launching a new climate program for certified sustainable products

from Techmeme https://ift.tt/32Ulkib
via A.I .Kung Fu

California looks to ban gas guzzlers – but legal hurdles abound

California Governor Gavin Newsom made a bold attempt today to ban sales of new gas-guzzling cars and trucks, marking a critical step in the state’s quest to become carbon neutral by 2045. But the effort to clean up the state’s largest source of climate emissions is almost certain to face serious legal challenges, particularly if President Donald Trump is re-elected in November.

Newsom issued an executive order that directs state agencies, including the California Air Resources Board, to develop regulations requiring every new passenger car and truck sold in the state to be zero-emissions vehicles by 2035. That pretty much limits future sales to electric vehicles (EVs) powered by batteries or hydrogen fuel cells. Similar rules would go into effect for most medium and heavy-duty vehicles by 2045.

If those rules are enacted, the roughly 2 million new vehicles sold in the state each year will all suddenly be EVs, providing a huge boost to the still nascent sector.

“California policy, especially automotive policy, has cascading effects across the US and even internationally, just because of the scale of our market,” says Alissa Kendall, professor of civil and environmental engineering at the University of California, Davis.

Indeed, the order would mean more auto companies will produce more EV lines, scaling up manufacturing and driving down costs. The growing market would, in turn, create greater incentives to build out the charging or hydrogen fueling infrastructure necessary to support it all.

The move also could make a big dent in transportation emissions. Passenger and heavy-duty vehicles together account for more than 35% of the state’s climate pollution, which has proven an especially tricky share to reduce in a sprawling state of car loving-residents (indeed, California’s vehicle emissions have been ticking up). 

But Newsom’s executive order only goes so far. It doesn’t address planes, trains, or ships, and it could take another couple decades for residents to stop driving all the gas-powered vehicles already on the road.

Whether the rules go into effect at all, and to what degree, will depend on many variables, including what legal grounds the Air Resources Board uses to justify the policies, says Danny Cullenward, a lecturer at Stanford’s law school focused on environmental policy.

One likely route is for the board to base the new regulations on tailpipe emissions standards, which California has used in the past to force automakers to produce more fuel-efficient vehicles, and nudge national standards forward. But that approach may require obtaining a new waiver from the Environmental Protection Agency allowing the state to exceed the federal government’s vehicle emissions rules under the Clean Air Act, the source of an already heated battle between the state and the Trump administration.

Last year, Trump announced he would revoke California’s earlier waiver to set tighter standards, prompting the state and New York to sue. So whether California can pursue this route could depend on how courts view the issue and who is sitting in the White House come late January.

It’s very likely that the automotive industry will challenge the rules no matter how the state goes about drafting them. And the outcome of those cases could depend on which court it lands in—and, perhaps eventually, who is sitting on the Supreme Court.

But whatever legal hurdles it may face, California and other states need to rapidly cut auto emissions to have any hope of combating the rising threat of climate change, says Dave Weiskopf, senior policy advisor with NextGen Policy in Sacramento.

“This is what science requires and it’s the next logical step for state policy,” he says.

from MIT Technology Review https://ift.tt/2G4b5Pr
via A.I .Kung Fu

DOJ proposes congressional fix of Section 230 as Trump turns up heat on Big Tech - CNET

Revisions to the law would weaken liability protections for social media companies that fact-check the president's posts.

from CNET News https://ift.tt/3hVu8Zu
via A.I .Kung Fu

Lordstown Motors claims it has 40,000 preorders, shows off its new interior - Roadshow

It's also close to becoming a publicly traded company, according to its representatives.

from CNET News https://ift.tt/2FXRMYj
via A.I .Kung Fu

2021 Kia Sorento X-Line is ready to hit the trails - Roadshow

The rugged X-Line trim level for the new Sorento crossover gets more ground clearance and different styling elements.

from CNET News https://ift.tt/303hkKo
via A.I .Kung Fu

2021 Kia Sorento X-Line has rugged looks and comes in a great green color - Roadshow

The new Sorento's X-Line package is only available on the top SX-Prestige trim level.

from CNET News https://ift.tt/2EsGcDH
via A.I .Kung Fu

Next iPad Pro will feature mini-LED displays, Apple analyst Kuo predicts - CNET

The company will reportedly speed up adoption of the displays in its iPad and MacBook lineups.

from CNET News https://ift.tt/3hVDTqr
via A.I .Kung Fu

Every new movie and show on Netflix: October 2020 - CNET

The Haunting of Bly Manor, Schitt's Creek, Rebecca and more!

from CNET News https://ift.tt/33RdPb7
via A.I .Kung Fu

Save $200 on Tineco's PureOne S12 Pro EX cordless stick vacuum - CNET

You get two removable batteries, a dual-charging wall mount and a stick vacuum that converts to a hand vac.

from CNET News https://ift.tt/3kD91g0
via A.I .Kung Fu

BMW announces its most extreme M model ever -- and it only costs $33,000 - Roadshow

The catch is that the M 1000RR is a motorcycle, but what a motorcycle it is.

from CNET News https://ift.tt/32YJXue
via A.I .Kung Fu

BMW M 1000RR brings insane, race-level performance to the street - Roadshow

Motorrad's new take on its long-lived and much-loved S 1000RR promises to give Ducati's Panigale V4 models a real run for their money.

from CNET News https://ift.tt/33XcnE7
via A.I .Kung Fu

AT&T 5G is coming to three Air Force bases - CNET

Buckley Air Force Base, Joint Base Elmendorf-Richardson and Offutt Air Force Base are getting AT&T 5G.

from CNET News https://ift.tt/2RUIQ8g
via A.I .Kung Fu

iPad Air 4 vs. iPad Mini (2019): Which is best for you?

With the announcement of the new Apple iPad Air 4, we compare it across six core areas with the iPad Mini to see which of these tablets you should pick up.

from Emerging Tech | Digital Trends https://ift.tt/3kHx0uB
via A.I .Kung Fu

California Plans to Ban Sales of Gas-Powered Cars by 2035

Governor Gavin Newsom outlines an ambitious plan for the nation's largest state to rely exclusively on electric-powered passenger cars and trucks.

from Wired https://ift.tt/32WfBrY
via A.I .Kung Fu

SEC filing: Peter Thiel and Richard Li have created a SPAC called Bridgetown Holdings Ltd., targeting SE Asia investment, and are seeking to raise $575M via IPO (Lizette Chapman/Bloomberg)

Lizette Chapman / Bloomberg:
SEC filing: Peter Thiel and Richard Li have created a SPAC called Bridgetown Holdings Ltd., targeting SE Asia investment, and are seeking to raise $575M via IPO  —  - Bridgetown also backed by Richard Li's Pacific Century  — SPAC aims to raise $575 million, according to filing

from Techmeme https://ift.tt/3mH2EKJ
via A.I .Kung Fu

CrowdStrike has agreed to acquire Preempt Security, a provider of zero trust and conditional access tech for enterprise threat management, for about $96M (Natalie Gagliordi/ZDNet)

Natalie Gagliordi / ZDNet:
CrowdStrike has agreed to acquire Preempt Security, a provider of zero trust and conditional access tech for enterprise threat management, for about $96M  —  The company said it plans to use the deal to bolster its Falcon platform with conditional access technology.

from Techmeme https://ift.tt/3crZbLg
via A.I .Kung Fu

Samsung unveils its Galaxy S20 FE, with a 6.5-inch display, Snapdragon 865, 120Hz refresh rate, starting at $699 for sub-6GHz model or $749 for mmWave 5G model (Dieter Bohn/The Verge)

Dieter Bohn / The Verge:
Samsung unveils its Galaxy S20 FE, with a 6.5-inch display, Snapdragon 865, 120Hz refresh rate, starting at $699 for sub-6GHz model or $749 for mmWave 5G model  —  The ‘Fan Edition’ will be available October 2nd  —  Samsung already makes three different kinds of Galaxy S20 phones: a regular, a Plus, and an Ultra.

from Techmeme https://ift.tt/33P010G
via A.I .Kung Fu

Sources: GoodRx, a price comparison app for prescription drugs at local pharmacies, raises $1.14B in its IPO at $33 per share, valuing the company ~$12.7B (Dan Primack/Axios)

Dan Primack / Axios:
Sources: GoodRx, a price comparison app for prescription drugs at local pharmacies, raises $1.14B in its IPO at $33 per share, valuing the company ~$12.7B  —  GoodRx, a price comparison app for prescription drugs at local pharmacies, on Tuesday raised $1.14 billion in its IPO, Axios has learned.

from Techmeme https://ift.tt/32QmQ4S
via A.I .Kung Fu

TransferWise reported annual revenues of £302.6M in FY ending March 2020, up 70% YoY, and net profit of £21.3M, says it has 8M customers globally, up 33% YoY (Alex Wilhelm/TechCrunch)

Alex Wilhelm / TechCrunch:
TransferWise reported annual revenues of £302.6M in FY ending March 2020, up 70% YoY, and net profit of £21.3M, says it has 8M customers globally, up 33% YoY  —  TransferWise, a European fintech unicorn, announced the financial results of its fiscal year ending March, 2020.

from Techmeme https://ift.tt/2EsObRi
via A.I .Kung Fu

Facebook removes accounts, groups, and Pages tied to China that posted about US politics; the network was followed by fewer than 3K US accounts (New York Times)

New York Times:
Facebook removes accounts, groups, and Pages tied to China that posted about US politics; the network was followed by fewer than 3K US accounts  —  The social media campaign was small but targeted all sides of the debate.  Officials said Beijing had not decided whether to wade more directly in the American presidential race.

from Techmeme https://ift.tt/33LPGTe
via A.I .Kung Fu

Morgan Beller, one of the founders of Facebook's Libra digital currency, has left the company to become a partner at VC firm NFX (Salvador Rodriguez/CNBC)

Salvador Rodriguez / CNBC:
Morgan Beller, one of the founders of Facebook's Libra digital currency, has left the company to become a partner at VC firm NFX  —  - One of the co-founders of Facebook's libra digital currency and Novi payments wallet has left the company before the release of either technology.

from Techmeme https://ift.tt/2G5va7J
via A.I .Kung Fu

2021 Hyundai Sonata N-Line looks ready to run laps and turn heads - Roadshow

The Sonata is the latest model to get the performance-minded N-Line upgrade, and we're thrilled about the results.

from CNET News https://ift.tt/3kSgNTJ
via A.I .Kung Fu

Hyundai's 2021 Sonata N-Line adds performance chops to its good looks - Roadshow

The Sonata was already easy on the eyes, but now it'll be a blast on a canyon road.

from CNET News https://ift.tt/360UTcG
via A.I .Kung Fu

You won't find a cheaper pet camera than the Petcube, now $36 - CNET

This budget camera sticks anywhere and includes an optional vet chat subscription service.

from CNET News https://ift.tt/3cmWXN0
via A.I .Kung Fu

Tesla Battery Day: New 'biscuit tin' battery cells could change everything - Roadshow

The new, larger cells will offer more range and less cost thanks to the integration of new production technology.

from CNET News https://ift.tt/2HocJvU
via A.I .Kung Fu

Tesla Model S Plaid announced with 520-mile range, 200-mph top speed and world-beating acceleration - Roadshow

The craziest Model S yet is available to order now with deliveries starting in late 2021.

from CNET News https://ift.tt/3kEejYF
via A.I .Kung Fu

UFC 253: Adesanya vs. Costa -- how to watch, start time and full fight card - CNET

Two flashy, undefeated rising stars fight for the UFC middleweight title at UFC 253.

from CNET News https://ift.tt/2ZZBLrA
via A.I .Kung Fu

An old TV crashed entire town's broadband every day for more than a year - CNET

Electrical interference from the appliance knocked out internet services in a small Welsh town.

from CNET News https://ift.tt/3hTMxG1
via A.I .Kung Fu

5G is coming to San Francisco's BART train system - CNET

The plan includes Wi-Fi on all trains, small cells along the rail line and 5G antennas in every underground station.

from CNET News https://ift.tt/2RMvN8W
via A.I .Kung Fu

Tesla's Elon Musk promises full self-driving Autopilot beta in 'a month or so' - Roadshow

Tesla's CEO shared that its engineers have fully overhauled the Autopilot software stack and are almost ready to share a dramatic upgrade.

from CNET News https://ift.tt/36c8ga7
via A.I .Kung Fu

2021 BMW M3 marks the return of the king, hopefully - Roadshow

Controversial grille and all, here's the new M3.

from CNET News https://ift.tt/3iVjFyt
via A.I .Kung Fu

2021 BMW M4 coupe is bigger, more powerful and uglier than before - Roadshow

The new BMW M4 has a standard manual transmission, up to 503 horsepower and available all-wheel drive.

from CNET News https://ift.tt/360tQOn
via A.I .Kung Fu

The 2021 BMW M4 coupe has that big grille and the craziest seats I've ever seen - Roadshow

BMW's new M4 also has up to 503 horsepower, optional all-wheel drive and tons of track-ready tech. But the seats are what really matters.

from CNET News https://ift.tt/3hXHAfc
via A.I .Kung Fu

2021 BMW M3 revealed: The OG sport sedan returns with 473 hp and a controversial face - Roadshow

Yep, it's got that grille, but it sounds like the M division has cooked up quite a machine this time around.

from CNET News https://ift.tt/3029xMP
via A.I .Kung Fu

ISS performs avoidance maneuver over 'unknown piece of space debris' - CNET

Space junk was coming too close for comfort to the International Space Station.

from CNET News https://ift.tt/33Q6ZCL
via A.I .Kung Fu

'Dayglow' at Rosetta's comet turns out to be unexpected ultraviolet aurora - CNET

Years after the Rosetta mission landed on Comet Chury, a new phenomenon has been excavated from its data.

from CNET News https://ift.tt/2ZWfZVC
via A.I .Kung Fu

NASA spacecraft discovers chunks of another asteroid on asteroid Bennu - CNET

The Osiris-Rex mission team spotted some Vesta crumbs on Bennu's face.

from CNET News https://ift.tt/2HpIQeD
via A.I .Kung Fu

The Swift Project announces its language tools are now available for Windows 10 (Joseph Keller/iMore)

Joseph Keller / iMore:
The Swift Project announces its language tools are now available for Windows 10  —  Swift is taking flight on Windows. … The Swift Project, the ongoing open-source efforts to develop the Swift programming language, has announced that an initial release of Swift is now available for Windows 10.

from Techmeme https://ift.tt/3iPLJ6h
via A.I .Kung Fu

Overview of the state of bundles in 2020 across Netflix, Disney, Amazon, Apple, and Microsoft (Ben Thompson/Stratechery)

Ben Thompson / Stratechery:
Overview of the state of bundles in 2020 across Netflix, Disney, Amazon, Apple, and Microsoft  —  If the famous Jim Barksdale quote is to be believed — the one about there only being two ways to make money in business, bundling and unbundling — then I am long past due for a follow-up to 2017's The Great Unbundling.

from Techmeme https://ift.tt/3iTiobe
via A.I .Kung Fu

Interview with Satya Nadella and Phil Spencer on buying more game companies, investing in Xbox Game Pass, and Microsoft's rejected TikTok bid (Ian Sherr/CNET)

Ian Sherr / CNET:
Interview with Satya Nadella and Phil Spencer on buying more game companies, investing in Xbox Game Pass, and Microsoft's rejected TikTok bid  —  In an interview, Microsoft CEO Satya Nadella and Xbox head Phil Spencer discuss further acquisitions, how to keep making hits and what happened with TikTok.

from Techmeme https://ift.tt/3iVZMqO
via A.I .Kung Fu