Connect with us

AI

Ex-Google policy chief dumps on the tech giant for dodging human rights

Avatar

Published

on

Google’s ex-head of international relations, Ross LaJeunesse — who clocked up more than a decade working government and policy-related roles for the tech giant before departing last year — has become the latest (former) Googler to lay into the company for falling short of its erstwhile “don’t be evil” corporate motto.

Worth noting right off the bat: LaJeunesse is making his own pitch to be elected as a U.S. senator for the Democrats in Maine, where he’s pitting himself against the sitting Republican, Susan Collins. So this lengthy blog post, in which he sets out reasons for joining (“making the world better and more equal”) and — at long last — exiting Google does look like an exercise in New Year reputation “exfoliation,” shall we say.

One that’s intended to anticipate and deflect any critical questions he may face on the campaign trail, given his many years of service to Mountain View. Hence the inclusion of overt political messaging, such as lines like: “No longer can massive tech companies like Google be permitted to operate relatively free from government oversight.”

Still, the post makes more awkward reading for Google. (Albeit, less awkward than the active employee activism the company continues to face over a range of issues — from its corporate culture and attitude toward diversity to product dev ethics.)

LaJeunesse claims that (unnamed) senior management actively evaded his attempts to push for it to adopt a company-wide Human Rights program that would, as he tells it, “publicly commit Google to adhere to human rights principles found in the UN Declaration of Human Rights, provide a mechanism for product and engineering teams to seek internal review of product design elements, and formalize the use of Human Rights Impact Assessments for all major product launches and market entries.”

“[E]ach time I recommended a Human Rights Program, senior executives came up with an excuse to say no,” LaJeunesse alleges, going on to claim that he was subsequently side-lined in policy discussions related to a censored search project Google had been working on to enable it to return to the Chinese market.

The controversial project, code-named Dragonfly, was later shut down, per LaJeunesse’s telling, after Congress raised questions — backing up the blog’s overarching theme that only political scrutiny can put meaningful limits on powerful technologists. (Check that already steady drumbeat for the 2020 U.S. elections.)

He writes:

At first, [Google senior executives] said human rights issues were better handled within the product teams, rather than starting a separate program. But the product teams weren’t trained to address human rights as part of their work. When I went back to senior executives to again argue for a program, they then claimed to be worried about increasing the company’s legal liability. We provided the opinion of outside experts who re-confirmed that these fears were unfounded. At this point, a colleague was suddenly re-assigned to lead the policy team discussions for Dragonfly. As someone who had consistently advocated for a human rights-based approach, I was being sidelined from the on-going conversations on whether to launch Dragonfly. I then realized that the company had never intended to incorporate human rights principles into its business and product decisions. Just when Google needed to double down on a commitment to human rights, it decided to instead chase bigger profits and an even higher stock price.

Reached for comment, a Google spokesman sent us this statement, attributed to a Google spokeswoman: “We have an unwavering commitment to supporting human rights organisations and efforts. That commitment is unrelated to and unaffected by the reorganisation of our policy team, which was widely reported and which impacted many members of the team. As part of this reorganisation, Ross was offered a new position at the exact same level and compensation, which he declined to accept. We wish Ross all the best with his political ambitions.”

LaJeunesse’s blog post also lays into Google’s workplace culture — making allegations that bullying and racist stereotyping were commonplace.

Including even apparently during attempts by management to actively engage with the issue of diversity…

It was no different in the workplace culture. Senior colleagues bullied and screamed at young women, causing them to cry at their desks. At an all-hands meeting, my boss said, “Now you Asians come to the microphone too. I know you don’t like to ask questions.” At a different all-hands meeting, the entire policy team was separated into various rooms and told to participate in a “diversity exercise” that placed me in a group labeled “homos” while participants shouted out stereotypes such as “effeminate” and “promiscuous.” Colleagues of color were forced to join groups called “Asians” and “Brown people” in other rooms nearby.

We’ve asked Google for comment on these allegations and will update this post with any response.

It’s clearly a sign of the “techlash” times that an ex-Googler, who’s now a senator-in-the-running, believes there’s political capital to be made by publicly unloading on his former employer. 

“The role of these companies in our daily lives, from how we run our elections to how we entertain and educate our children, is just too great to leave in the hands of executives who are accountable only to their controlling shareholders who — in the case of Google, Amazon, Facebook and Snap — happen to be fellow company insiders and founders,” LaJeunesse goes on to write, widening his attack to incorporate other FAANG giants.

Expect plenty more such tech giant piñata in the run up to November’s ballot.

Read more: https://techcrunch.com/2020/01/02/ex-google-policy-chief-dumps-on-the-tech-giant-for-dodging-human-rights/

AI

HyperRec: Efficient Recommender Systems with Hyperdimensional Computing

Avatar

Published

on

Home

TECHNICAL PAPERS

Brain-inspired computing model. The new algorithm, called HyperRec, uses data that is modeled with binary vectors in a high dimension.

popularity

A group of researchers are taking a different approach to AI.

The University of California at San Diego, the University of California at Irvine, San Diego State University and DGIST recently presented a paper on a new hardware algorithm based on hyperdimensional (HD) computing, which is a brain-inspired computing model. The new algorithm, called HyperRec, uses data that is modeled with binary vectors in a high dimension.

Technical paper link is here.

Coinsmart. Beste Bitcoin-Börse in Europa
Source: https://semiengineering.com/hyperrec-efficient-recommender-systems-with-hyperdimensional-computing/

Continue Reading

AI

Build a scalable machine learning pipeline for ultra-high resolution medical images using Amazon SageMaker

Avatar

Published

on

Neural networks have proven effective at solving complex computer vision tasks such as object detection, image similarity, and classification. With the evolution of low-cost GPUs, the computational cost of building and deploying a neural network has drastically reduced. However, most techniques are designed to handle pixel resolutions commonly found in visual media. For example, typical resolution sizes are 544 and 416 pixels for YOLOv3, 300 and 512 pixels for SSD, and 224 pixels for VGG. Training a classifier over a dataset consisting of gigapixel images (10^9+ pixels) such as satellite or digital pathology images is computationally challenging. These images cannot be directly input into a neural network because each GPU is limited by available memory. This requires specific preprocessing techniques such as tiling to be able to process the original images in smaller chunks. Furthermore, due to the large size of these images, the overall training time tends to be high, often requiring several days or weeks without the use of proper scaling techniques such as distributed training.

In this post, we explain how to build a highly scalable machine learning (ML) pipeline to fulfill three objectives:

Dataset

In this post, we use a dataset consisting of whole-slide digital pathology images obtained from The Cancer Genome Atlas (TCGA) to accurately and automatically classify them as LUAD (adenocarcinoma), LUSC (squamous cell carcinoma), or normal lung tissue, where LUAD and LUSC are the two most prevalent subtypes of lung cancer. The dataset is available for public use by NIH and NCI.

The raw high-resolution images are in SVS format. SVS files are used for archiving and analyzing Aperio microscope images. You can apply the techniques and tools used in this post to any ultra high-resolution image dataset, including satellite images.

The following is a sample image of a tissue slide. This single image contains over a quarter of a billion pixels, and occupies over 750 MB of memory. This image cannot be fed directly to a neural network in its original form, so we must tile the image into many smaller images.

The following are samples of tiled images generated after preprocessing the preceding tissue slide image. These RGB 3-channel images are of size 512×512 and can be directly used as inputs to a neural network. Each of these tiled images is assigned the same label as the parent slide. Additionally, tiled images with more than 50% background are discarded.

Architecture overview

The following figure shows the overall end-to-end architecture, from the original raw images to inference. First, we use SageMaker Processing to tile, zoom, and sort the images into train and test splits, and then package them into the necessary number of shards for distributed SageMaker training. Second, a SageMaker training job loads the Docker container from Amazon Elastic Container Registry (Amazon ECR). The job uses Pipe mode to read the data from the prepared shards of images, trains the model, and stores the final model artifact in Amazon Simple Storage Service (Amazon S3). Finally, we deploy the trained model on a real-time inference endpoint that loads the appropriate Docker container (from Amazon ECR) and model (from Amazon S3) to process inference requests with low latency.

Data preprocessing using SageMaker Processing

The SVS slide images are preprocessed in three steps:

  • Tiling images – The images are tiled by non-overlapping 512×512 pixel windows, and tiles containing over 50% background are discarded. The tiles are stored as JPEG images.
  • Converting images to TFRecords – We use SageMaker Pipe mode to reduce our training time, which requires the data to be available in a proto-buffer format. TFRecord is a popular proto-buffer format used for training models with TensorFlow. We explain SageMaker Pipe mode and proto-buffer format in more detail in the following section.
  • Sorting TFRecords – We sort the dataset into test, train, and validation cohorts for a three-way classifier (LUAD/LUSC/Normal). The TCGA dataset can have multiple slide images corresponding to a single patient. We need to make sure all the tiles generated from slides corresponding to the same patient occupy the same split to avoid data leakage. For the test set, we create per-slide TFRecord containing all the tiles from that slide so that we can evaluate the model in the way it will be used in deployment.

The following is the preprocessing code:

def generate_tf_records(base_folder, input_files, output_file, n_image, slide=None): record_file = output_file count = n_image with tf.io.TFRecordWriter(record_file) as writer: while count: filename, label = random.choice(input_files) temp_img = plt.imread(os.path.join(base_folder, filename)) if temp_img.shape != (512, 512, 3): continue count -= 1 image_string = np.float32(temp_img).tobytes() slide_string = slide.encode('utf-8') if slide else None tf_example = image_example(image_string, label, slide_string) writer.write(tf_example.SerializeToString())

We use SageMaker Processing for the preceding preprocessing steps, which allows us to run data preprocessing or postprocessing, feature engineering, data validation, and model evaluation workloads with SageMaker. Processing jobs accept data from Amazon S3 as input and store processed output data back into Amazon S3.

A benefit of using SageMaker Processing is the ease of distributing inputs across multiple compute instances. We can simply set s3_data_distribution_type=ShardedByS3Key parameter to divide data equally among all processing containers.

Importantly, the number of processing instances matches the number of GPUs we will use for distributed training with Horovod (i.e., 16). The reasoning becomes clearer when we introduce Horovod training.

The processing script is available on GitHub.

processor = Processor(image_uri=image_name, role=get_execution_role(), instance_count=16, # run the job on 16 instances base_job_name='processing-base', # should be unique name instance_type='ml.m5.4xlarge', volume_size_in_gb=1000) processor.run(inputs=[ProcessingInput( source=f's3://<bucket_name>/tcga-svs', # s3 input prefix s3_data_type='S3Prefix', s3_input_mode='File', s3_data_distribution_type='ShardedByS3Key', # Split the data across instances destination='/opt/ml/processing/input')], # local path on the container outputs=[ProcessingOutput( source='/opt/ml/processing/output', # local output path on the container destination=f's3://<bucket_name>/tcga-svs-tfrecords/' # output s3 location )], arguments=['10000'], # number of tiled images per TF record for training dataset wait=True, logs=True)

Distributed model training using SageMaker Training

Taking ML models from conceptualization to production is typically complex and time-consuming. We have to manage large amounts of data to train the model, choose the best algorithm for training it, manage the compute capacity while training it, and then deploy the model into a production environment. SageMaker reduces this complexity by making it much easier to build and deploy ML models. It manages the underlying infrastructure to train your model at petabyte scale and deploy it to production.

After we preprocess the whole-slide images, we still have hundreds of gigabytes of data. Training on a single instance (GPU or CPU) would take several days or weeks to finish. To speed things up, we need to distribute the workload of training a model across multiple instances. For this post, we focus on distributed deep learning based on data parallelism using Horovod, a distributed training framework, and SageMaker Pipe mode.

Horovod: A cross-platform distributed training framework

When training a model with a large amount of data, the data needs to distributed across multiple CPUs or GPUs on either a single instance or multiple instances. Deep learning frameworks provide their own methods to support distributed training. Horovod is a popular framework-agnostic toolkit for distributed deep learning. It utilizes an allreduce algorithm for fast distributed training (compared with a parameter server approach) and includes multiple optimization methods to make distributed training faster. For more examples of distributed training with Horovod on SageMaker, see Multi-GPU and distributed training using Horovod in Amazon SageMaker Pipe mode and Reducing training time with Apache MXNet and Horovod on Amazon SageMaker.

SageMaker Pipe mode

You can provide input to SageMaker in either File mode or Pipe mode. In File mode, the input files are copied to the training instance. With Pipe mode, the dataset is streamed directly to your training instances. This means that the training jobs start sooner, compute and download can happen in parallel, and less disk space is required. Therefore, we recommend Pipe mode for large datasets.

SageMaker Pipe mode requires data to be in a protocol buffer format. Protocol buffers are language-neutral, platform-neutral, extensible mechanisms for serializing structured data. TFRecord is a popular proto-buffer format used for training models with TensorFlow. TFRecords are optimized for use with TensorFlow in multiple ways. First, they make it easy to combine multiple datasets and integrate seamlessly with the data import and preprocessing functionality provided by the library. Second, you can store sequence data—for instance, a time series or word encodings—in a way that allows for very efficient and (from a coding perspective) convenient import of this type of data.

The following diagram illustrates data access with Pipe mode.

Data sharding with SageMaker Pipe mode

You should keep in mind a few considerations when working with SageMaker Pipe mode and Horovod:

  • The data that is streamed through each pipe is mutually exclusive of the other pipes. The number of pipes dictates the number of data shards that need to be created.
  • Horovod wraps the training script for each compute instance. This means that data for each compute instance needs to be from a different shard.
  • With the SageMaker Training parameter S3DataDistributionType set to ShardedByS3Key, we can share a pipe with more than one instance. The data is streamed in round-robin fashion across instances.

To illustrate this better, let’s say we use two instances (A and B) of type ml.p3.8xlarge. Each ml.p3.8xlarge instance has four GPUs. We create four pipes (P1, P2, P3, and P4) and set S3DataDistributionType = 'ShardedByS3Key’. As shown in the following table, each pipe equally distributes the data between two instances in a round-robin fashion. This is the core concept needed in setting up pipes with Horovod. Because Horovod wraps the training script for each GPU, we need to create as many pipes as there are GPUs per training instance.


The following code shards the data in Amazon S3 for each pipe. Each shard should have a separate prefix in Amazon S3.

# Definite distributed training hyperparameters
train_instance_type='ml.p3.8xlarge'
train_instance_count = 4
gpus_per_host = 4
num_of_shards = gpus_per_host * train_instance_count distributions = {'mpi': { 'enabled': True, 'processes_per_host': gpus_per_host }
}

# Sharding
client = boto3.client('s3')
result = client.list_objects(Bucket=s3://<bucket_name>, Prefix='tcga-svs-tfrecords/train/', Delimiter='/') j = -1
for i in range(num_of_shards): copy_source = { 'Bucket': s3://<bucket_name>, 'Key': result['Contents'][i]['Key'] } print(result['Contents'][i]['Key']) if i % gpus_per_host == 0: j += 1 dest = 'tcga-svs-tfrecords/train_sharded/' + str(j) +'/' + result['Contents'][i]['Key'].split('/')[2] print(dest) s3.meta.client.copy(copy_source, s3://<bucket_name>, dest) # Define inputs to SageMaker estimator
svs_tf_sharded = f's3://<bucket_name>/tcga-svs-tfrecords'
shuffle_config = sagemaker.session.ShuffleConfig(234)
train_s3_uri_prefix = svs_tf_sharded
remote_inputs = {} for idx in range(gpus_per_host): train_s3_uri = f'{train_s3_uri_prefix}/train_sharded/{idx}/' train_s3_input = s3_input(train_s3_uri, distribution ='ShardedByS3Key', shuffle_config=shuffle_config) remote_inputs[f'train_{idx}'] = train_s3_input remote_inputs['valid_{}'.format(idx)] = '{}/valid'.format(svs_tf_sharded)
remote_inputs['test'] = '{}/test'.format(svs_tf_sharded)
remote_inputs

We use a SageMaker estimator to launch training on four instances of ml.p3.8xlarge. Each instance has four GPUs. Thus, there are a total of 16 GPUs. See the following code:

local_hyperparameters = {'epochs': 5, 'batch-size' : 16, 'num-train':160000, 'num-val':8192, 'num-test':8192} estimator_dist = TensorFlow(base_job_name='svs-horovod-cloud-pipe', entry_point='src/train.py', role=role, framework_version='2.1.0', py_version='py3', distribution=distributions, volume_size=1024, hyperparameters=local_hyperparameters, output_path=f's3://<bucket_name>/output/', instance_count=4, instance_type=train_instance_type, input_mode='Pipe') estimator_dist.fit(remote_inputs, wait=True)

The following code snippet of the training script shows how to orchestrate Horovod with TensorFlow for distributed training:

mpi = False
if 'sagemaker_mpi_enabled' in args.fw_params: if args.fw_params['sagemaker_mpi_enabled']: import horovod.keras as hvd mpi = True # Horovod: initialize Horovod. hvd.init() # Pin GPU to be used to process local rank (one GPU per process) gpus = tf.config.experimental.list_physical_devices('GPU') tf.config.experimental.set_visible_devices(gpus[hvd.local_rank()], 'GPU')
else: hvd = None callbacks = []
if mpi: callbacks.append(hvd.callbacks.BroadcastGlobalVariablesCallback(0)) callbacks.append(hvd.callbacks.MetricAverageCallback()) if hvd.rank() == 0: callbacks.append(ModelCheckpoint(args.output_dir + '/checkpoint-{epoch}.ckpt', save_weights_only=True, verbose=2))
else: callbacks.append(ModelCheckpoint(args.output_dir + '/checkpoint-{epoch}.ckpt', save_weights_only=True, verbose=2)) train_dataset = train_input_fn(hvd, mpi)
valid_dataset = valid_input_fn(hvd, mpi)
test_dataset = test_input_fn()
model = model_def(args.learning_rate, mpi, hvd)
logging.info("Starting training")
size = 1
if mpi: size = hvd.size() model.fit(train_dataset, steps_per_epoch=((args.num_train // args.batch_size) // size), epochs=args.epochs, validation_data=valid_dataset, validation_steps=((args.num_val // args.batch_size) // size), callbacks=callbacks, verbose=2)

Because Pipe mode streams the data to each of our instances, the training script cannot calculate the data size during training (which is needed to compute steps_per_epoch). The parameter is therefore provided manually as a hyperparameter to the TensorFlow estimator. Additionally, the number of data points must be specified so that it can be divided equally amongst the GPUs. An unequal division could lead to a Horovod deadlock, because the time taken by each GPU to complete the training process is no longer identical. To ensure that the data points are equally divided, we use the same of number of instances for preprocessing as the number of GPUs for training. In our example, this number is 16.

Inference and deployment

After we train the model using SageMaker, we deploy it for inference on new images. To set up a persistent endpoint to get one prediction at a time, use SageMaker hosting services. To get predictions for an entire dataset, use SageMaker batch transform.

In this post, we deploy the trained model as a SageMaker endpoint. The following code deploys the model to an m4 instance, reads tiled image data from TFRecords, and generates a slide-level prediction:

# Generate predictor object from trained model
predictor = estimator_dist.deploy(initial_instance_count=1, instance_type='ml.m4.xlarge') # Tile-level prediction
raw_image_dataset = tf.data.TFRecordDataset(f'images/{local_file}') # read a TFrecord
parsed_image_dataset = raw_image_dataset.map(dataset_parser) # Parse TFrecord to JPEGs pred_scores_list = []
for i, element in enumerate(parsed_image_dataset): image = element[0].numpy() label = element[1].numpy() slide = element[2].numpy().decode() if i == 0: print(f"Making tile-level predictions for slide: {slide}...") print(f"Querying endpoint for a prediction for tile {i+1}...") pred_scores = predictor.predict(np.expand_dims(image, axis=0))['predictions'][0] pred_class = np.argmax(pred_scores) if i > 0 and i % 10 == 0: plt.figure() plt.title(f'Tile {i} prediction: {pred_class}') plt.imshow(image / 255) pred_scores_list.append(pred_scores)
print("Done.") # Slide-level prediction (average score over all tiles)
mean_pred_scores = np.mean(np.vstack(pred_scores_list), axis=0)
mean_pred_class = np.argmax(mean_pred_scores)
print(f"Slide-level prediction for {slide}:", mean_pred_class)

The model is trained on individual tile images. During inference, the SageMaker endpoint provides classification scores for each tile. These scores are averaged out across all tiles to generate the slide-level score and prediction. The following diagram illustrates this workflow.

A majority vote scheme would also be appropriate.

To perform inference on a large new batch of slide images, you can run a batch transform job for offline predictions on the dataset in Amazon S3 on multiple instances. Once the processed TFRecords are retrieved from Amazon S3, you can replicate the preceding steps to generate a slide-level classification for each of the new images.

Conclusion

In this post, we introduced a scalable machine learning pipeline for ultra high-resolution images that uses SageMaker Processing, SageMaker Pipe mode, and Horovod. The pipeline simplifies the convoluted process of large-scale training of a classifier over a dataset consisting of images that approach the gigapixel scale. With SageMaker and Horovod, we eased the process by distributing inputs across multiple compute instances, which reduces training time. We also provided a simple but effective strategy to aggregate tile-level predictions to produce slide-level inference.

For more information about SageMaker, see Build, train, and deploy a machine learning model with Amazon SageMaker. For the complete example to run on SageMaker, in which Pipe mode and Horovod are applied together, see the GitHub repo.

References

  1. Nicolas Coudray, Paolo Santiago Ocampo, Theodore Sakellaropoulos, Navneet Narula, Matija Snuderl, David Fenyö, Andre L. Moreira, Narges Razavian, Aristotelis Tsirigos. “Classification and mutation prediction from non–small cell lung cancer histopathology images using deep learning”. Nature Medicine, 2018; DOI: 10.1038/s41591-018-0177-5
  2. https://github.com/ncoudray/DeepPATH/tree/master/DeepPATH_code
  3. https://www.cancer.gov/about-nci/organization/ccg/research/structural-genomics/tcga

About the Authors

Karan Sindwani is a Data Scientist at Amazon Machine Learning Solutions where he builds and deploys deep learning models. He specializes in the area of computer vision. In his spare time, he enjoys hiking.

Vinay Hanumaiah is a Deep Learning Architect at Amazon ML Solutions Lab, where he helps customers build AI and ML solutions to accelerate their business challenges. Prior to this, he contributed to the launch of AWS DeepLens and Amazon Personalize. In his spare time, he enjoys time with his family and is an avid rock climber.

Ryan Brand is a Data Scientist in the Amazon Machine Learning Solutions Lab. He has specific experience in applying machine learning to problems in healthcare and the life sciences, and in his free time he enjoys reading history and science fiction.

Tatsuya Arai, Ph.D. is a biomedical engineer turned deep learning data scientist on the Amazon Machine Learning Solutions Lab team. He believes in the true democratization of AI and that the power of AI shouldn’t be exclusive to computer scientists or mathematicians.

Coinsmart. Beste Bitcoin-Börse in Europa
Source: https://aws.amazon.com/blogs/machine-learning/building-a-scalable-machine-learning-pipeline-for-ultra-high-resolution-medical-images-using-amazon-sagemaker/

Continue Reading

AI

Apply to join Transform’s annual Tech Showcase

Avatar

Published

on

Join Transform 2021 this July 12-16. Register for the AI event of the year.


VentureBeat’s annual Tech Showcase returns at Transform 2021: Accelerating Enterprise Transformation with AI and Data, hosted July 12-16.

VentureBeat will be selecting 5 companies to present the latest and greatest AI and data products on the main stage during the first day of Transform 2021.

We are looking for companies that have built the coolest products and solutions leveraging bleeding edge technologies to help businesses achieve real and tangible results using AI and data. Whether you are a stealth startup or Fortune 500, or anywhere in between, we welcome your submission.

If you have a story to tell, and an AI or Data product/solution with tangible business results and demonstrative use cases, please submit your application here before 5pm PST on Tuesday, June 1.

Selected companies will present on Transform’s main stage in front of hundreds of industry decision makers and will be featured in our on-demand video-hub following the event.

Attendees from across the globe will join online to hear from top industry experts on strategy and technology in the main application areas of AI/ML automation technology, data, analytics, intelligent automation, conversational AI, intelligent AI assistants, AI at the edge, IoT, & computer vision. Executives across industries are invited to join Transform 2021, so register today to join VentureBeat for 5-days of AI and data.

QUESTIONS

Contact events@venturebeat.com with any questions regarding your application.

VentureBeat

VentureBeat’s mission is to be a digital town square for technical decision-makers to gain knowledge about transformative technology and transact. Our site delivers essential information on data technologies and strategies to guide you as you lead your organizations. We invite you to become a member of our community, to access:

  • up-to-date information on the subjects of interest to you
  • our newsletters
  • gated thought-leader content and discounted access to our prized events, such as Transform 2021: Learn More
  • networking features, and more

Become a member

Coinsmart. Beste Bitcoin-Börse in Europa
Source: https://venturebeat.com/2021/05/12/apply-to-join-transforms-annual-tech-showcase/

Continue Reading

AI

TikTok removes 500k+ accounts in Italy after DPA order to block underage users

Avatar

Published

on

Video sharing social network TikTok has removed more than 500,000 accounts in Italy following an intervention by the country’s data protection watchdog earlier this year ordering it to recheck the age of all Italian users and block access to any under the age of 13.

Between February 9 and April 21 more than 12.5M Italian users were asked to confirm that they are over 13 years old, according to the regulator.

Online age verification remains a hard problem and it’s not clear how many of the removed accounts definitively belonged to under 13s. The regulator said today that TikTok removed over 500k users because they were “likely” to be under the age of 16; around 400,000 because they declared an age under 13 and 140,000 through what the DPA describes as “a combination of moderation and reporting tools” implemented within the app.

TikTok has also agreed to take a series of additional measures to strengthen its ability to detect and block underage users — including potentially developing AI tools to help it identify when children are using the service.

Reached for comment, TikTok sent us a statement confirming that it is trialling “additional measures to help ensure that only users aged 13 or over are able to use TikTok”.

Here’s the statement, which TikTok attributed to Alexandra Evans, its head of child safety in Europe:

“TikTok’s top priority is protecting the privacy and safety of our users, and in particular our younger users. Following continued engagement with the Garante, we will be trialling additional measures to help ensure that only users aged 13 or over are able to use TikTok.

“We already take industry-leading steps to promote youth safety on TikTok such as setting accounts to private by default for users aged under 16 and enabling parents to link their account to their teen’s through Family Pairing. There is no finish line when it comes to safety, and we continue to evaluate and improve our policies, processes and systems, and consult with external experts.”

Italy’s data protection regulator made an emergency intervention in January — ordering TikTok to recheck the age of all users and block any users whose age it could not verify. The action followed reports in local media about a 10-year-old girl from Palermo who died of asphyxiation after participating in a “blackout challenge” on the social network.

Among the beefed up measures TikTok has agreed to take is a commitment to act faster to remove underage users — with the Italian DPA saying the platform has guaranteed it will cancel reported accounts it verifies as belonging to under 13s within 48 hours.

The regulator said TikTok has also committed to “study and develop” solutions — which may include the use of artificial intelligence — to “minimize the risk of children under 13 using the service”.

TikTok has also agree to launch ad campaigns, both in app and through radio and newspapers in Italy, to raise awareness about safe use of the platform and get the message out that it is not suitable for under-12s — including targeting this messaging in a language and format that’s likely to engage underage minors themselves.

The social network has also agreed to share information with the regulator relating to the effectiveness of the various experimental measures — to work with the regulator to identify the best ways of keeping underage users off the service.

The DPA said it will continue to monitor TikTok’s compliance with its commitments.

Prior to the Garante’s action, TikTok’s age verification checks had been widely criticized as trivially easier for kids to circumvent — with children merely needing to input a false birth date that suggested they are older than 13 to circumvent the age gate and access the service.

A wider investigation that the DPA opened into TikTok’s handling and processing of children’s data last year remains ongoing.

The regulator announced it had begun proceedings against the platform in December 2020, following months of investigation, saying then that it believed TikTok was not complying with EU data protection rules which set stringent requirements for processing children’s data.

In January the Garante also called for the European Data Protection Board to set up an EU taskforce to investigate concerns about the risks of children’s use of the platform — highlighting similar concerns being raised by other agencies in Europe and the U.S.

In February the European consumer rights organization, BEUC, also filed a series of complaints against TikTok, including in relation to its handling of kids’ data.

Earlier this year TikTok announced plans to bring in outside experts in the region to help with content moderation and said it would open a ‘transparency’ center in Europe where outside experts could get information on its content, security and privacy policies.

Coinsmart. Beste Bitcoin-Börse in Europa
Source: https://techcrunch.com/2021/05/12/tiktok-removes-500k-accounts-in-italy-after-dpa-order-to-block-underage-users/

Continue Reading
Aviation5 days ago

What Happened To Lufthansa’s Boeing 707 Aircraft?

Blockchain5 days ago

Launch of Crypto Trading Team by Goldman Sachs

Aviation4 days ago

JetBlue Hits Back At Eastern Airlines On Ecuador Flights

Aviation5 days ago

United Airlines Uses The Crisis To Diversify Latin American Network

Cyber Security4 days ago

Cybersecurity Degrees in Massachusetts — Your Guide to Choosing a School

Blockchain5 days ago

Miten tekoälyä käytetään videopeleissä ja mitä tulevaisuudessa on odotettavissa

Blockchain5 days ago

DOGE Co-founder Reveals the Reasons Behind its Price Rise

Blockchain4 days ago

“Privacy is a ‘Privilege’ that Users Ought to Cherish”: Elena Nadoliksi

Cyber Security5 days ago

U.S. and the U.K. Published Attack on IT Management Company SolarWinds

AI1 day ago

Build a cognitive search and a health knowledge graph using AWS AI services

Fintech5 days ago

The Spanish fintech Pecunpay strengthens its position as a leader in the issuance of corporate programs

Blockchain News5 days ago

Nasdaq-Listed Metromile Backs Bitcoin for its Insurance Products

Blockchain5 days ago

Crypto.com Now Enabling Cosmos’ Inter-Blockchain Communication Functionality for Cross-Chain Transfers

Cyber Security4 days ago

Cybersecurity Degrees in Texas — Your Guide to Choosing a School

SaaS5 days ago

Blockchain23 hours ago

Meme Coins Craze Attracting Money Behind Fall of Bitcoin

Cleantech5 days ago

What We Know About Tesla’s “Bobcat Project”

Energy2 days ago

ONE Gas to Participate in American Gas Association Financial Forum

Aviation5 days ago

Ryanair Goes Full Steam Ahead On Portugal Capacity Expansion

Blockchain4 days ago

Opimas estimates that over US$190 billion worth of Bitcoin is currently at risk due to subpar safekeeping

Trending