Connect with us


NLP in 2020; Modern Applications




NLP has gone from rule based systems to generative systems with almost human level accuracy along multiple rubrics within 40 years. This is incredible considering we were so far off naturally talking to a computer system even just ten years ago; now I can tell Google Home to turn off my sitting room lights.

In the Stanford Lecture by Chris Manning introduces a Computer Science class to what NLP is, its complexity and specific toolings such as word2vec which enable learning systems to learn from natural language. Professor Manning is the Director of the Stanford Artificial Intelligence Laboratory and is a leader in applying Deep Learning (DL) to NLP.

The goal of NLP is to allow computers to ‘understand’ natural language in order to perform tasks and support the human user to make decisions. For a logic system, understanding and representing the meaning of language is a “difficult goal”. The goal is so compelling all major technology firms have put huge investment into the field. The lecture focuses on these areas of the NLP challenge.

Some applications which you might encounter NLP systems are spell checking, search, recommendations, speech recognition, dialog agents, sentiment analysis and translation services. One key point Chris Manning explains is that human language (either text, speech or movement) is unique in that it is done to communicate something, some ‘meaning’ is embedded in the action. This is not often the case with anything else that generates data. Its data with intent, extracting and understanding the intent is part of the NLP challenge. Chris Manning also lists “Why NLP is hard” which I think we take for granted.

Language interpretation depends on ‘common sense’ and contextual knowledge, language is ambiguous (computers like direct, formal statements!), language contains a complex mix of situational, visual and linguistic knowledge from various timelines. Learning systems we have now do not have a lifetime of learned weights and bias so can only currently be applied in narrow-AI use cases.

The Stanford lecture also dives into DL and how it is different to a human exploring and designing features or signals to then apply to the learning systems. The lecture discusses the first spark of DL with speech recognition from work done by George Dahl and how the DL approach got a 33% increase in performance compared to traditional feature modelling. Professor Manning also talks about how NLP and DL have added capabilities in three segments, namely what he calls Levels; speech, words, syntax and semantics. Tools; parts-of-speech, entities and parsing and Applications; machine translation, sentiment analysis, dialogue agent and question answering. Stating NLP + DL have created a ‘Few key tools’ which have wide applications.

Words as vectors —

Towards the end of the lecture we explore the ideas around how words are represented as numbers in vector spaces and how this applies to NLP and DL. Word meaning vectors then are usable to represent meaning in words, sentences and beyond.



How AI is Changing The Game Of Business




Artificial intelligence (AI) is changing the game of business at an astonishing rate.


1. 8 Proven Ways to Use Chatbots for Marketing (with Real Examples)

2. How to Use Texthero to Prepare a Text-based Dataset for Your NLP Project

3. 5 Top Tips For Human-Centred Chatbot Design

4. Chatbot Conference Online


Continue Reading


Artificial Intelligence and Online Privacy: Blessing and a Curse




Author profile picture

Artificial Intelligence (AI) is a beautiful piece of technology made to seamlessly augment our everyday experience. It is widely utilized in everything starting from marketing to even traffic light moderation in cities like Pittsburg. However, swords have two edges and the AI is no different. There are a fair number of upsides as well as downsides that follow such technological advancements.

One way or another, the technology is moving too quickly while the education about the risks and safeguards that are in place are falling behind for the vast majority of the population. The whole situation is as much of a blessing for humankind as it is a curse.

In this article, we will be mainly discussing how the AI is being utilized, how the ease of processing data-enabled companies and government agencies to have power over online privacy, and how to stay careful of possible abuses of said power.

A Blessing

Have you ever noticed that every application lately has been asking your permission for personalized ads? These are advertisements from companies that are targeting you. However, how do they know what you may want? This is due to the hard work done by a good number of AIs. They are constantly sifting through and analyzing what you are looking at. Everything is being calculated like: for how long or what you are even messaging sometimes.

Do not be afraid though, they are not reading your secret messages. Reputable companies have these robots looking for some keywords in your conversation. For example, if you are speaking to your friend about bicycles the AI gets excited and offers you an ad from Specialized, Cube, or Ghost offering their brand new mountain bikes. This is how companies like Facebook, Google, Twitter, and others make money. If the product is free – you are the product. In these cases, a product for marketing companies that pay hefty amounts to have their ads on one or all of these platforms.

AI in Everyday Life | The Good

The AI enhances user experience to totally new levels. Unfortunately, analyzing and sifting through huge masses of data is not something the human brain specializes in. To properly process a big chunk of data the companies have to hire a team of individuals. The whole ordeal can take hours, days, weeks, or even months depending on what they are trying to analyze and we are not even taking into consideration the financial expenses.

Even with the professionals, there are moments of subjective bias involved in the result. Humans also tend to flat out ignore or forget a good number of details. An AI though is a different case. Data processing is their specialization. It is something they love doing the most and are happy to continue doing until the end of time. The process is becoming more and more efficient the longer this process goes on. The whole point of artificial intelligence is the ability to enhance their capabilities by finding better and efficient ways to do the same task.

Banking, for example, is the one that has benefitted the most from an AI. As an example, most of the top banks now offer online financial advising. This is done by AIs on the starting level. These programs analyze your spending, give visualized charts, and help save money where possible.

Different exchanges like stock market, foreign exchange market (Forex, FX), cryptocurrency markets, etc. are using the AIs to predict where the market is going to move. Apart from this, the historic data is also always being summarized for personal and corporate use. As it was once said the future is nothing but a repetition of the past.

AI in Military | The Good

The military is very fond of the potential of AI. In 2018 the US announced that they will be implementing artificial intelligence in every branch where it is applicable. The AI is going to help the military in analyzing intelligence, making lightning-fast decisions, automatization of vehicles, weaponry, and logistics as a whole. It is worth noting that the US military is one of the prime places to develop such cutting edge development since they are showering in finances. In 2020 the US congress has approved as much as $718 billion for the Department of Defense out of which $927 million is going to be moving towards the AI and machine learning development.

It is worth noting that the technology has evolved quite a lot during the last decade. Although the good nature of this is debatable the idea that no military personnel has to risk their lives to gather information that can be done using a robot is a huge achievement.

AI in Cyber Security | The Good

Although technically doing cyber crimes is harder now since there are a bunch of protocols available to protect end-user from possible threats it’s also easier because there are a fair number of scripts available that automate very tiresome and complex processes.

Cybersecurity is related to several problems that experts have been trying to deal with as they go along. Once the corporate network becomes large the surface at which the attacks may proceed increases exponentially. This can be done either via breaching application security, network security, or basic social engineering. Today’s systems are protected using multilayer security systems. During the development, the software undergoes an intense testing phase to root out all of the possible hoops that hackers can use to exploit a vulnerability. After the product is finished, there are teams working day and night on enhancing, adding functionality, and patching previous problems. Once the code becomes big enough though the issues are extremely hard to cover and the IT teams have to prioritize some over the others. This is why some of the bugs in your favorite video games are not being patched out quickly and may stay forever if they aren’t game-breaking.

Companies usually build their infrastructure with their intranet protocols that are hiding all of the necessary information from unwelcome guests. However, this means that the network security aspect has to also be covered fully. This is done via firewalls, anti-virus and anti-malware software, data loss prevention, virtual private networks, etc. A good number of programs are written to alleviate this whole process. For example, behavioral analysis is almost solely done with artificial intelligence. This means that there is a software setting and gathering data about how certain individuals are acting while they are inside of the system. The moment there is a suspicious activity the AI either has the right to restrict access temporarily or straight up direct an IT specialist to take a closer look.

A Curse

Artificial Intelligence is a double-edged sword though. It is important to understand that there are a fair number of precautions that need to be taken before utilizing a service of your choice. It is important to know which company is pulling the strings. Your personal information can and will be utilized to gain more and more funds. However, some companies are going to tell you upfront and some won’t. National security agencies have been targeting these data-gathering companies for years already. Google for example has much more capacities to spy on people than a lot of developed, developing, and underdeveloped countries. The data that goes through their servers is a huge point of interest in a number of both good and bad people. Some of the companies are going to monetize on these offers selling out their collected data to the highest bidder. 

AI for Data Collection | The Bad

The European Union has enacted a General Data Protection Regulation (GDPR) law for companies that are operating in the EU space. The GDPR regulates the acquisition of information from online customers. Numerous methods are utilized to achieve this goal. It can be anything from what we have already mentioned like robots skimming through the data that you freely share (like status updates, likes, etc.) or more hidden files designated for this very same purpose like cookies that you download from websites upon visit. There are different types of cookies, of course, but all of them can be utilized to gather data. This is why currently most of the websites are asking for your permission to use them. This is why it is recommended to check the cookie policy for a website that you frequently visit. For the ones that you don’t – just don’t allow cookies at all.

Most of the time cookies are used to enhance already existing websites. They are the sole purpose why it takes considerably less time to load pages once you’ve already visited them. However, some of these cookies (tracker cookies) are used to track your activity on the internet. This is not even limited to the website that you are using. For example, Facebook cookies get activated every time there is a company API being used. This means that every page that uses Facebook comments, likes, or other services automatically activate their cookies on your computer. This is why you can be commenting something on one website with Twitter addon and then get an ad centered around that very same topic on Twitter. Although I have not mentioned the AI here it is obvious that all of these processes are going through machine learning algorithms that process and dish out the results.

You cannot fully hide your activities on the internet though. People have utilized virtual machines to do different tasks, turned off all of the data collection options, used VPN, but the core idea is that the trail, even though very convoluted at this point, is still going to be leading towards you and it all depends on how badly someone or something wants to find you. Either way, you need to trust someone when you are using the internet.

Why Trust One Company Over the Other?

This leads us to one of the final topics of this article. Why should you trust specific developer companies and not others? 

Apart from the general reputation that company gets it’s important to understand that where the firm is based and takes orders from is a big deciding factor. For example, why are Chinese companies not trusted overal? There is a fair argument saying that if you don’t use TikTok due to concerns about Chinese agencies collecting your data you shouldn’t be using Instagram because the US national security departments will also be doing the same. However, things are not that straightforward. 

Apple vs FBI | A Case of Integrity

There have been cases where these agencies have demanded access to user data or even straight up requested to have a backdoor into every device. Apple’s dispute against the FBI in 2016 is a prime example of one such instance. To summarize in 2015 the FBI extracted iPhone 5C from a shooter who participated in a December 2015 San Bernardino terrorist attack. They demanded Apple to unlock this device and Tim Cook, the CEO of Apple, has kindly rejected this notion. The company has outlined the importance of preventing terrorism but they couldn’t just give out a backdoor to every iPhone to the FBI. This is in light of the situation that the US government, or the National Security Agency (NSA) in particular, was already outed to be spying on the citizens by a famous whistleblower and founder of WikiLeaks, Edward Snowden.

Apple took this case to court where they used an argument that their customer privacy would be in vain if such instance comes to happen. The FBI didn’t just want one iPhone to be unlocked but they wanted a software backdoor into every past, current, and future smartphone from Apple. The court has ruled in favor of Apple and the case was closed.

This is a happy ending to this scenario. However, we have to keep in mind that the US constitution gives companies an ability to argue such cases in court. Whereas if such thing happens in China, Russia, or any less-democratic country the companies simply do not have the power to go to court and defend their case. 

Company Origin Matters

The political nature of the nation is also extremely important. All of the middle to large-sized corporations in China are under the direct supervision of the Chinese Communist Party and share the information with their national security agencies. 

Tencent, for example, which is a Chinese company that has developed most of the popular applications like TikTok, WeChat, and has numerous companies under its umbrella-like Riot Games (League of Legends, Valorant, Teamfight Tactics, Legends of Runterra) and Epic Games does not and cannot uphold such standards. More importantly, the CEO and founder father if the company Ma Huateng is a prominent member of the CCP as well as one of the most influential persons across the world. The argument can be made that TikTok doesn’t keep US customer data on Chinese servers but this is a soft barrier. A company can transfer this data for as much as they like and there are a good number of triggers that the Chinese government can pull to make them do their bidding.

A department of this very same company called Tencent Keen Security Lab has researched the security features of Tesla car’s autopilot function. They messed around with sensors and found ways to make the AI go haywire. Tencent is also a 5% stakeholder of Tesla company. This is important information to keep in mind.


The 21st century is a period of technological wonders. Things that were only available in SciFi movies and books are slowly coming to life. While technology is integrating itself into our daily lives it is important to start learning about security aspects as well as the technology that is being utilized. Being picky is not bad when it comes to the usage of the internet. It is strongly encouraged by security professionals across the world. While the digital world may seem like it’s separate from our real lives. The reality stands that it is an integral part and as real as the physical one.

The usage of an AI is making it possible to enlarge operations to scales unseen before. There are some positive sides like ad campaigns but also negative ones like surveillance systems working against consumers. This is all dependent on who’s driving the car and we, the users, can decide which cars to ride and who will be the driver. 

This article is designated to highlight some of the uses of AI and how it augments our lives. The core idea is that if users become knowledgable about the subject we can pick and choose which companies to trust. In a capitalist market, consumer trust drives income and the companies will act in any way necessary to keep their user bases.


The Noonification banner

Subscribe to get your daily round-up of top tech stories!


Continue Reading


Introducing Our Low Code Machine Learning Platform




Author profile picture

@kfugereKyle Fugere

Head of Innovation & Ventures

We are very excited to release the free tier of dunnhumby Model Lab this as part of our partnership with Microsoft. dunnhumby Model Lab is an application that provides automated pipelines for deploying machine learning algorithms and has been used to build millions of models on behalf of our clients.

Sign up here:

We make it easy to connect your data, clean your data, and run your machine learning pipeline within minutes. You can then take that output and copy right into a notebook for further refinement if needed.


Run all your machine learning from a single platform. 

You can create new projects, reference datasets, and create
multiple experiments in just a few clicks!

You can also follow the progress of your machine learning
experiments as they update in real-time.

Automated Tuning of Machine Learning Algorithms

Machine learning algorithms have many parameters that need
to be calibrated based on the data being used.

Some algorithms have over 15 parameters that need to be tuned in order to operate properly, all of them having an impact on each other. It’s billions of combinations. If done by hand, this process can take days without any guarantee you will find the one that will give you the best results.

To reduce the time-to-value and allow data scientists to focus on the good models, Model Lab provides a built-in module that automatically tunes your machine learning algorithms.

Leveraging state-of-the-art Bayesian optimization, Model Lab can tune any machine learning algorithm in a fraction of the time usually required.

Parallel Computing & Resource Optimization

We leverage Kubernetes to run all the models in parallel. This results in a significant boost in performance and considerable reduction in runtime and, therefore, time-to-value.

Each model runs as a container on our cluster, allowing multiple models to be trained simultaneously.

Model Lab comes with a resource optimization module that optimizes
the amount of RAM allocated to each container, allowing us to train as many models as possible in parallel.

See our Medium article on this work here!

Our Experiments

Build a classification predictive model in minutes

Classification is one of the most common types of predictive models done at dunnhumby and across our clients. It has been used to predict things like retention insurance, customer churn for retailers and even loyalty.

Building a classification model has become mainstream nowadays and clients expect results very quickly. However, many steps must be completed before delivering a predictive model, which gets in the way of delivering results quickly when performed manually.

Classification is Model Lab’s machine learning experiment that automates the end-to-end process of building a simple but strong classification predictive model. Originally designed to deliver preliminary results within minutes to validate the data, project scope and hypothesis, FastLog has also proven many times to be at par with more complex machine learning algorithms in term of performance for production purposes.

Build a high-level view of your data

Clustering is one of the most common type of analysis done at dunnhumby. It has been used to group stores, products, and customers based on loyalty and lifestyles, with unique behavior, which is different from the respective in other groups.

Clustering has become mainstream at dunnhumby hence, clients expect quick results with interpretations. A lot of steps and methods needs to
be tried before getting the best result in clustering analysis.

Clustering is Model Lab’s experiment for clustering. It automates the end-to-end process of building the best clustering model using given data and very few parameters. FastCluster can perform multiple clustering iteration and identify the best results very quickly.

Data cleaner – Get your data ready to get to work

Cleaner is a utility solution from Model Lab that aims at quickly getting your data ready for the work by cleaning and making it ready. It is a requisite for all our experiments.

Most of the time, the raw data is not in a state that can run machine learning algorithms. Things like missing values, characters, duplicated
rows, etc… can take up to 60% of data scientists time and is the least


Coming Soon

Classification – Multiclass

This is an extension of the current Classification experiment to support multi-class problems.


This experiment will be able to predict a continuous target.

Driver Analysis / Non-linear

We will be releasing a new version of our Driver Analysis experiment that leverages non-linear algorithms. Those machine learning algorithms have an advantage over traditional methodologies like univariate analysis, in the sense that they explore both non-linear relationships and interactions in the data.

3D Data Exploration

Data visualization techniques have proven to be sometimes very useful to identify pattern in the data, as our brain is very good a finding patterns. This module will leverage PCA and t-SNE techniques, and provide a 3D visualization of the projected data.

Time Series Modelling

Many problems look at the evolution of certain metrics or target over time. This experiment will allow users to run such analysis and make forecasting over time.

This product is the brainchild of Dr. Victor Robin and is part of dunnhumby Labs, dunnhumby’s new product accelerator.


The Noonification banner

Subscribe to get your daily round-up of top tech stories!


Continue Reading
CoinTelegraph22 mins ago

Inside the blockchain developer’s mind: The upgradeability crisis

CoinTelegraph28 mins ago

Uniswap 57% price crash marks ‘DeFi depression’ for SUSHI, CRV, UNI

CoinTelegraph34 mins ago

DeFi and security not mutually exclusive, say CertiK and Binance Smart Chain Reps

Biotechnology1 hour ago

LabCorp to roll out Resolution Bio’s lung cancer liquid biopsy test

Biotechnology2 hours ago

Silverback Therapeutics nets $85M top-up to push lead ADC through phase 1

Biotechnology2 hours ago

Cognoa’s AI-powered diagnostic app for autism clears clinical study, heads for FDA review

Cannabis2 hours ago

Cannabis Investment Basics: Debt v. Equity

CoinTelegraph2 hours ago

Huobi crypto exchange launches trading app in Russia

Blockchain3 hours ago

Cointelegraph Consulting survey: Mass DeFi adoption is 3–10 years away

Biotechnology4 hours ago

FDA to set high bar for emergency COVID-19 vaccine approvals

Biotechnology4 hours ago

Roche, AC Immune’s tau-blocking drug flops in Alzheimer’s as biotech’s shares halved

Big Data4 hours ago

5 Must-Know Google Analytics Strategies For SEO Success

Big Data4 hours ago

How Cryptocurrency Is Benefiting From Big Data Analytics

Blockchain4 hours ago

Despite $2,000 Drop, Bitcoin Is Holding a Critical Macro Level

Private Equity5 hours ago

KKR buys into Reliance Retail at $57bn valuation as firms scramble to back business

Private Equity6 hours ago

RedBird Capital seals N3 sale to Accenture as businesses scramble for digital sales expertise amid Covid-19

Business Insider6 hours ago

Johnson & Johnson Initiates Pivotal Global Phase 3 Clinical Trial of Janssen’s COVID-19 Vaccine Candidate

Business Insider6 hours ago

Veteran Industry Auditor Chris Johnson Joins Comscore to Bolster Measurement Standards Compliance Initiatives as Comscore Expands Efforts with the Media Rating Council

Business Insider6 hours ago

$15.7 Billion Mobile Device Management Market – Global Forecast to 2025 with Use Cases from IBM, Blackberry, Microsoft, 42Gears Mobility Systems and Soti

Business Insider6 hours ago

RedBird Capital Agrees To Sell N3 To Accenture; Terms Not Disclosed – Quick Facts

Private Equity6 hours ago

Palladium picks up century-old modular steel buildings, hardware business Trachte USA

Private Equity6 hours ago

These 5 private equity billionaries just made it into Forbes’ 400 richest Americans list

Blockchain6 hours ago

The Bears to Follow the Crypto Market But Bulls to Outpower Soon

Blockchain6 hours ago

After 50% Drop, Chainlink (LINK) Is Reaching a Do or Die Technical Level

Techcrunch6 hours ago

Salsify nabs $155M as its commerce experience platform sees a big surge of business from Covid-19

Private Equity7 hours ago

Hg seals trade sale exit of A-Plan Group after more than doubling revenue at company

Techcrunch8 hours ago

New research shows European startups are spending drastically less on a US launch, for the same gains

Blockchain8 hours ago

Bitcoin is Eerily Copying Gold Trends; Analyst Fears Breakdown

Startups8 hours ago

‘AI-powered’ fitness app Freeletics scores $25M Series B

Techcrunch9 hours ago

Apple partner Servify raises $23 million to scale its devices after-sales and management platform overseas

Automotive9 hours ago

VW’s Traton Group strikes a deal with TuSimple to develop autonomous trucks

Entrepreneur10 hours ago

Melorra Secures 12.5M USD from Symphony Asia and Others

Fintech10 hours ago

RegTech: An opportunity to boost Australia and UK trade relations?

CNBC10 hours ago

Nokia says it’s focused on tech and customers, but political fights complicate things

Blockchain10 hours ago

3 Crucial Reasons Why Bitcoin Could Fall Below $10K This Fall

Fintech10 hours ago

What is Ethereum?

Entrepreneur10 hours ago

BYJU’s Gets Undisclosed Fund From BlackRock and Others

Blockchain11 hours ago

Wirex Announces its First Crowdfunding and Aiming to Raise £1M

CNBC12 hours ago

Tesla CEO Elon Musk touts new tech at ‘Battery Day,’ offers 2020 delivery guidance

CNBC12 hours ago

Jamie Dimon says he’s OK with higher taxes on the rich, but wealth tax is ‘almost impossible’