Machine learning *

Hi All,

I have some good news for you…

Data Science Digest is back! We’ve been “offline” for a while, but no worries — You’ll receive regular digest updates with top news and resources on AI/ML/DS every Wednesday, starting today.

If you’re more used to getting updates every day, follow us on social media:

Telegram - https://t.me/DataScienceDigest
Twitter - https://twitter.com/Data_Digest
LinkedIn - https://www.linkedin.com/company/data-science-digest/
Facebook - https://www.facebook.com/DataScienceDigest/

And finally, your feedback is very much appreciated. Feel free to share any ideas with me and the team, and we’ll do our best to make Data Science Digest a better place for all.

Regards,
Dmitry Spodarets.

Author: Sergey Lukyanchikov, Sales Engineer at InterSystems

What is Distributed Artificial Intelligence (DAI)?

Attempts to find a “bullet-proof” definition have not produced result: it seems like the term is slightly “ahead of time”. Still, we can analyze semantically the term itself – deriving that distributed artificial intelligence is the same AI (see our effort to suggest an “applied” definition) though partitioned across several computers that are not clustered together (neither data-wise, nor via applications, not by providing access to particular computers in principle). I.e., ideally, distributed artificial intelligence should be arranged in such a way that none of the computers participating in that “distribution” have direct access to data nor applications of another computer: the only alternative becomes transmission of data samples and executable scripts via “transparent” messaging. Any deviations from that ideal should lead to an advent of “partially distributed artificial intelligence” – an example being distributed data with a central application server. Or its inverse. One way or the other, we obtain as a result a set of “federated” models (i.e., either models trained each on their own data sources, or each trained by their own algorithms, or “both at once”).

Distributed AI scenarios “for the masses”

We will not be discussing edge computations, confidential data operators, scattered mobile searches, or similar fascinating yet not the most consciously and wide-applied (not at this moment) scenarios. We will be much “closer to life” if, for instance, we consider the following scenario (its detailed demo can and should be watched here): a company runs a production-level AI/ML solution, the quality of its functioning is being systematically checked by an external data scientist (i.e., an expert that is not an employee of the company). For a number of reasons, the company cannot grant the data scientist access to the solution but it can send him a sample of records from a required table following a schedule or a particular event (for example, termination of a training session for one or several models by the solution). With that we assume, that the data scientist owns some version of the AI/ML mechanisms already integrated in the production-level solution that the company is running – and it is likely that they are being developed, improved, and adapted to concrete use cases of that concrete company, by the data scientist himself. Deployment of those mechanisms into the running solution, monitoring of their functioning, and other lifecycle aspects are being handled by a data engineer (the company employee).

There is a lot of commotion in text-to-speech now. There is a great variety of toolkits, a plethora of commercial APIs from GAFA companies (based both on new and older technologies). There are also a lot of Silicon Valley startups trying to ship products akin to "deep fakes" in speech.

But despite all this ruckus we have not yet seen open solutions that would fulfill all of these criteria:

• Naturally sounding speech;
• A large library of voices in many languages;
• Support for 16kHz and 8kHz out of the box;
• No GPUs / ML engineering team / training required;
• Unique voices not infringing upon third-party licenses;
• High throughput on slow hardware. Decent performance on one CPU thread;
• Minimalism and lack of dependencies. One-line usage, no builds or coding in C++ required;
• Positioned as a solution, not yet another toolkit / compilation of models developed by other people;
• Not affiliated by any means with ecosystems of Google / Yandex / Sberbank;

We decided to share our open non-commercial solution that fits all of these criteria with the community. Since we have published the whole pipeline we do not focus much on cherry picked examples and we encourage you to visit our project GitHub repo to test our TTS for yourself.

One of the most time-consuming steps while implementing a SIEM solution is writing and tuning "Playbook" – a set of reaction procedures SOC Team has to follow in case of alert triggering.

So during one of our projects we stoped for a moment and thought: "How can we optimize (ideally automate) the Playbook?"

There’s a lot of talk about machine learning nowadays. A big topic – but, for a lot of people, covered by this terrible layer of mystery. Like black magic – the chosen ones’ art, above the mere mortal for sure. One keeps hearing the words “numpy”, “pandas”, “scikit-learn” - and looking each up produces an equivalent of a three-tome work in documentation.

I’d like to shatter some of this mystery today. Let’s do some machine learning, find some patterns in our data – perhaps even make some predictions. With good old Python only – no 2-gigabyte library, and no arcane knowledge needed beforehand.

Interested? Come join us.

Neural networks (NN) and computer vision models in particular are known to perform well in specific tasks, but often fail to generalize to tasks they have not been trained on. A model that performs well on a food data may perform poorly on satellite images. 

A new model from OpenAI named CLIP claims to close this gap by a large margin. The paper Open AI wrote presenting CLIP demonstrates how the model may be used on a various classification datasets in a zero-shot manner. 

In this article, I will explain the key ideas of the model they proposed and show you the code to use it. 

Currently, there are hardly any high quality / modern / free / public voice activity detectors except for WebRTC Voice Activity Detector (link). WebRTC though starts to show its age and it suffers from many false positives.

Also in some cases it is crucial to be able to anonymize large-scale spoken corpora (i.e. remove personal data). Typically personal data is considered to be private / sensitive if it contains (i) a name (ii) some private ID. Name recognition is a highly subjective matter and it depends on locale and business case, but Voice Activity and Number Detection are quite general tasks.

Key features:

• Modern, portable;
• Low memory footprint;
• Superior metrics to WebRTC;
• Trained on huge spoken corpora and noise / sound libraries;
• Slower than WebRTC, but fast enough for IOT / edge / mobile applications;
• Unlike WebRTC (which mostly tells silence from voice), our VAD can tell voice from noise / music / silence;
• PyTorch (JIT) and ONNX checkpoints;

Typical use cases:

• Spoken corpora anonymization;
• Can be used together with WebRTC;
• Voice activity detection for IOT / edge / mobile use cases;
• Data cleaning and preparation, number and voice detection in general;
• PyTorch and ONNX can be used with a wide variety of deployment options and backends in mind;

Every time when the essential question arises, whether to upgrade the cards in the server room or not, I look through similar articles and watch such videos.

Channel with the aforementioned video is very underestimated, but the author does not deal with ML. In general, when analyzing comparisons of accelerators for ML, several things usually catch your eye:

• The authors usually take into account only the "adequacy" for the market of new cards in the United States;
• The ratings are far from the people and are made on very standard networks (which is probably good overall) without details;
• The popular mantra to train more and more gigantic models makes its own adjustments to the comparison;

The answer to the question "which card is better?" is not rocket science: Cards of the 20* series didn't get much popularity, while the 1080 Ti from Avito (Russian craigslist) still are very attractive (and, oddly enough, don't get cheaper, probably for this reason).

All this is fine and dandy and the standard benchmarks are unlikely to lie too much, but recently I learned about the existence of Multi-Instance-GPU technology for A100 video cards and native support for TF32 for Ampere devices and I got the idea to share my experience of the real testing cards on the Ampere architecture (3090 and A100). In this short note, I will try to answer the questions:

• Is the upgrade to Ampere worth it? (spoiler for the impatient — yes);
• Are the A100 worth the money (spoiler — in general — no);
• Are there any cases when the A100 is still interesting (spoiler — yes);
• Is MIG technology useful (spoiler — yes, but for inference and for very specific cases for training);

With ML projects still on the rise we are yet to see integrated solutions in almost every device around us. The need for processing power, memory and experimentation has led to machine learning and DL frameworks targeting desktop computers first. However once trained, a model may be executed in a more constrained environment on a smartphone or on an IoT device. A particularly interesting environment to run the model on is browser. Browser-based solutions may be used on a wide range of devices, desktop and mobile, online and offline. The topic of this post is how to prepare a model for the in-browser usage.

This post presents an end-to-end implementations of a model creation in Python and Node.js. The end goal is to create a model and to use it in a browser. I'll use TensorFlow and TensorFlow.js as main frameworks. One could train a model in Python and convert it to JS. Alternative is to train a model directly in javascript, hence omitting the conversion step.

I have more experience in Python and use it in my everyday work. I occasionally use javascript, but have very little experience in the contemporary front-end development. My hope from this post that python developers with little JS experience could use it to kick start their JS usage.

Challenges of real-time AI/ML computations

• A “high-load” customer portal is integrated with an online recommendation system. The plan is to reconfigure promo campaigns at the level of the entire retail network (we will assume that instead of a “flat” promo campaign master there will be used a “segment-tactic” matrix). What will happen to the recommender mechanisms? What will happen to data feeds and updates into the recommender mechanisms (the volume of input data having increased 25000 times)? What will happen to recommendation rule generation setup (the need to reduce 1000 times the recommendation rule filtering threshold due to a thousandfold increase of the volume and “assortment” of the rules generated)?
• An equipment health monitoring system uses “manual” data sample feeds. Now it is connected to a SCADA system that transmits thousands of process parameter readings each second. What will happen to the monitoring system (will it be able to handle equipment health monitoring on a second-by-second basis)? What will happen once the input data receives a new bloc of several hundreds of columns with data sensor readings recently implemented in the SCADA system (will it be necessary, and for how long, to shut down the monitoring system to integrate the new sensor data in the analysis)?
• A complex of AI/ML mechanisms (recommendation, monitoring, forecasting) depend on each other’s results. How many man-hours will it take every month to adapt those AI/ML mechanisms’ functioning to changes in the input data? What is the overall “delay” in supporting business decision making by the AI/ML mechanisms (the refresh frequency of supporting information against the feed frequency of new input data)?

We are proud to announce that we have built from ground up and released our high-quality (i.e. on par with premium Google models) speech-to-text Models for the following languages:

• English;
• German;
• Spanish;

You can find all of our models in our repository together with examples, quality and performance benchmarks. Also we invested some time into making our models as accessible as possible — you can try our examples as well as PyTorch, ONNX, TensorFlow checkpoints. You can also load our model via TorchHub.

	PyTorch	ONNX	TensorFlow	Quality	Colab
English (en_v1)	✓	✓	✓	link
German (de_v1)	✓	✓	✓	link
Spanish (es_v1)	✓	✓	✓	link

Общая информация

Разработка апплета для фото на документы самостоятельно

1. Подготовка

1.1 Добавьте репозиторий Maven Huawei в файл на уровне проекта build.gradle

This is a fourth article in the series of works (see also first one, second one, and third one) describing Machine Learning system based on Lattice Theory named 'VKF-system'. The program uses Markov chain algorithms to generate causes of the target property through computing random subset of similarities between some subsets of training objects. This article describes bitset representations of objects to compute these similarities as bit-wise multiplications of corresponding encodings. Objects with discrete attributes require some technique from Formal Concept Analysis. The case of objects with continuous attributes asks for logistic regression, entropy-based separation of their ranges into subintervals, and a presentation corresponding to the convex envelope for subintervals those similarity is computed.

This is a third article in the series of works (see also first one and second one) describing Machine Learning system based on Lattice Theory named 'VKF-system'. It uses structural (lattice theoretic) approach to representing training objects and their fragments considered to be causes of the target property. The system computes these fragments as similarities between some subsets of training objects. There exists the algebraic theory for such representations, called Formal Concept Analysis (FCA). However the system uses randomized algorithms to remove drawbacks of the unrestricted approach. The details follow…

When Alex Garland’s series Devs (on FX and Hulu) came out this year, it gave developers their own sexy Hollywood workup. Who knew that coders could get snarled into murder plots and love triangles just for designing machine learning programs? Or that their software would cause a philosophical crisis? Sure, the average day of a developer is more code writing than murder but what a thrill to author powerful new program.