Autonomy is quickly becoming the driving force of the automobile industry. Leading autonomous car makers – GM, Waymo, Tesla, Ford, etc. – are all aggressively improving the infrastructure necessary for autonomy. As there are multiple autonomous vehicle technologies at play, a hot discussion is going on in the market about which one is better – vision based (camera) or sensors based (LiDAR, radar, and ultrasonic).

Currently, pretty much all active players in the market, with the exception of Tesla, are moving forth with LiDAR. Tesla, on the other hand, is betting solely on cameras/vision.

Before we get into thick of the discussion, let’s quickly go through these automotive vehicle technologies (LiDAR, radar, and camera).

LiDAR

LiDAR (Light Detection and Ranging) devices are basically active sensors that emit high-frequency laser signals in quick succession (up to 150k pulses/sec). It measures the time taken for each signal to bounce back and calculate the distance between the vehicle and obstacles with high precision. The main concerns with LiDAR are that it is way too costly and doesn’t recognize colors and traffic signs.

Radar

LiDAR and Radar are essentially the same technologies operating with light waves at different frequencies. Radar emits low-frequency signals; therefore, produce less accurate outcomes (see the image below). While cost is not an issue with radar technology, being a sensor, it also doesn’t recognize colors and signs, which are crucial for autonomous driving.

Ultrasonic is another automotive vehicle technology in wide use today. Unlike LiDAR and radar, these sensors use soundwaves to measure the distance from obstacles.

Camera

Being a vision-based device, camera estimates the distance based on its relative size in the captured frame. The main advantage with camera is that it can recognize traffic signs and traffic light colors, which sensors cannot. While cost is another advantage with camera, but given the amount of efforts the goes into training camera-based systems to achieve sensor level accuracy, it is not a very cost-effective solution either.

Related read: Computer Vision Trends 2019

Now that we have brushed up the basics of main autonomous vehicle technologies, let’s see how they are faring in the real world.

Current Industry State

It is clear that one technology – be it vision or sensors – cannot provide the ultimate solution for driverless cars. As a senior analyst at Navigant puts in “Layering sensors with different capabilities, rather than just relying on a purely vision-based system, is ultimately a safer and more robust solution”.

So, the question here is, which one should be used as the primary autonomous vehicle technology, and which one as the secondary? As mentioned earlier, LiDAR is the industry’s first choice for driverless cars. Main reason being, cameras estimate the distance between objects based on relative size; however, LiDAR sensors know it with high precision.

However, Tesla’s CEO Elon Musk, who has long criticized LiDAR, recently called it “a fool’s errand” at Tesla’s Autonomy Day and claimed that everyone will drop it in the near future. As absurd as it sounds, Musk backed his claim with some reasoning (although it may sound unreasonable to most). First, LiDAR is costly hardware. Secondly, he demonstrated how camera is getting better and better at achieving sensor-level accuracy.

Currently, cost is undeniably the biggest let-down with LiDAR; however, all active players are putting an enormous amount of efforts to bring its cost down. For instance, in 2017, Google-backed Waymo claimed that it will build its own LiDAR sensors and will reduce the cost from $75,000 to $7,500 apiece. In March 2019, Waymo updated on its claim that it will soon start selling its in-house LiDAR devices to third-parties.

Another drawback, LiDAR becomes inaccurate in fog, rain, and snow, as it’s high-frequency signals detects these small particles and include them in the rendering. On the other hand, radar works quite effectively in such weather conditions and it also costs considerably less; therefore, is arguably a better substitute than LiDAR when used as secondary autonomous vehicle technology in conjunction with cameras.

Also Read: The Current State of AI in the Corporate World

What Future Holds for Autonomous Vehicle Technology

Given the ongoing developments, it is likely that the future will unfold either of the following two possibilities:

1. Camera-based vision systems achieve LiDAR level accuracy
2. LiDAR becomes cheaper than the cost of improving camera systems

Which will happen first is hard to say with any certainty. LiDAR prices have dropped over the last few years, so has the camera’s image processing. Tesla, with its Fleet Learning capabilities, is betting high on cameras and argues that its systems are designed to replicate human behavior and recognize the world around with colors and signs, not by distance. However, discarding LiDAR completely is also not an option at the current stage. So, the next few years will be crucial in deciding where the course of autonomous vehicles turns.

Looking for quality labeling of your fleet data?Connect with us

Moneyball was an eye-opener.

Today, it is a widely popular theory. But unfortunately, widely misunderstood too.

Most people watch the movie or read the book and say, “Wow, we should use more of this analytics to solve our business problems”. And that’s where they are wrong.

Moneyball theory isn’t about the importance of using analytics. It is about how you use it.

Analytics was being used in baseball for decades. What Billy Beane’s Oakland Athletics did differently in the 2002 season was that they went against the collective wisdom of baseball scouts and analytics, made decisions based on previously overlooked metrics and went on to become one of the most successful teams of the season despite being one of the most underpaid ones.

Let’s dig deeper into it to figure out what businesses can learn from the Moneyball theory and how they can use analytics in a way that would allow them to experience growth beyond what conventional business wisdom would.

Understand the Problem (Really understand it)

If you have seen the movie, there’s a very interesting and very tense round table discussion between Billy Beane and a bunch of experienced baseball scouts about ‘what the problem is?’ Here is the clip:

The point here is – in general, our industry experience helps us make better business decisions. But it also influences how we see and interpret a given problem. And sometimes it influences us to the point where it prevents us from seeing the problem from a different angle even when the previously tried and tested methods don’t seem to work.

Evaluating a given problem from different angles allows us to see different possible solutions – some of which can be better than the original one we are inclined to stick. In terms of analytics, it allows us to consider previously ignored metrics, which might be the key to the solution for that particular problem.

Trust Your Data (More than your guts)

The challenge with new, unorthodox solutions (like the one in Moneyball) is that quite often it would appear absurd or too risky. So, it is likely that you might find yourself in a situation where you experience, peers, and even your guts would tell you to go otherwise. But if you can see the clear logic in what your data suggests, then just stick with it. It will pay off.

When you are changing something, especially a well-established legacy process or system, you’re going to face the resistance. Not everyone will understand its potential, let alone accept it. The important thing is to be right and stick with it until you deliver results. Because sometimes the best way to convince people is to show them not to tell them.

Involve Leadership (Because data guys don’t make decisions)

In his Forbes article on Moneyball, Florian Zettelmeyer of Kellogg School of Management wrote: “Moneyball succeeded for the Oakland A’s not because of data analytics but because of Beane, the leader who understood the analytics’ potential and changed the organization so it could deliver on that potential.”

So, it is imperative that leaders sit down with the data teams, brainstorm with them, and use their industry expertise to assist them in coming up with the best possible strategies. Leadership’s involvement will also mitigate the resistance towards change and will accelerate the process of making analytics a part of the organization’s DNA.

To wrap things up, here are two images showing Oakland A’s standings in the 2002 (Moneyball) season:

How they stood salary-wise: 

How they stood performance-wise:

Have you seen the Moneyball movie or read the book? Do you think we missed something? Let us know in the comment section below. We will add it.

Outperform your competition with the power of analyticsConnect with us

At this point, most of us tech people know we need some data science & machine learning skills in order to survive and thrive. But with so much buzz around these technologies, it is easy to lose track of what’s really important.

I see a lot of people running after the hype, trying to acquire lots of me-too skills without due diligence. As a result, the mismatch between acquired and required skills continues to grow.

Amid this overwhelming mess, knowing what trends are currently shaping the data science and machine learning landscape can help you be prudent about identifying the right skills where you need to invest your time and efforts.

So, let’s see 5 such key data science and machine learning career trends for 2019 that will help you build a future-proof career.

1. Companies Want Specialization

‘Data Scientist’ and ‘Machine Learning Engineer’ are fascinating job titles, but they are incomplete. Today, the industry has matured and companies are looking for specializations within these fields. For instance, here is a glimpse of all the data roles at Netflix.

Besides, most AI startups operate in niches; therefore, require skills specific to that niche. For example, if a company is building an NLP solution, instead of posting a job vacancy for ‘machine learning engineers’ it would look for ‘NLP engineers’. In fact, if you look it up on LinkedIn, you will see every machine learning or data science job vacancy with some sort of specialization with it.

2. The Demand of Data Engineers Rises

There is an oft-cited LinkedIn survey that states that in 2018, the demand for data engineers exceeds the demand for data scientists. Data engineers are responsible for developing software components of analytics applications. They collect and store data and do real-time processing to ensure uninterrupted data flow so that data scientists can analyze it seamlessly.

For the past couple of years, companies have been hiring data scientists relentlessly. As a result, now they don’t have enough resources to provide their data scientists with the required infrastructure, which automatically makes data engineering one of the most prominent digital skills to have in 2019.
Apart from being proficient in programming, data engineers also need to be proficient in Hadoop, MapReduce, Hive, MySQL, Cassandra, MongoDB, NoSQL, SQL, & Data streaming and programming.

3. Industry Drastically Lacks AIOps Engineers

Quick definition: AIOps for data science/machine learning solutions is what DevOps is for traditional software development.

The abundance of data scientists (only in comparison, the industry still needs a lot of data scientists) has not only increased the demand of data engineers, but it has also triggered the demand of engineers at the deployment end (AIOPs). The rough chart below sheds more light on it.

If we consider the current state of AI in the corporate world, the industry has enough resources focused on training models. But every model needs regular data and model versioning at the deployment end to ensure that the model continues to meet a business’s dynamic demand. And there, the industry faces a drastic lack of resources currently. So, all in all, this area currently presents heaps of opportunities. And the technology stack you need to learn for that includes frameworks such as TensorFlow Serving, Docker, Kubernetes (K8s), and Kubeflow.

4. Python Is the Present & Future

On the internet, there are already tons of resources on Python Vs R Vs SAS. But when it comes to machine learning and data science, it is already established (although arguably) that Python is the way to go, because it has the packages specifically designed for these jobs.

For beginners the trouble is that lots of tutorials and courses on the internet are based on R. For instance, on e-learning platform Data Camp, roughly 2/3rd of data science and machine learning tutorials are based on R, only 1/3rd in Python. But if you look at their respective communities, the Python community exceeds the R community by a great margin.

Now, I am not recommending that you don’t learn R at all. It is useful for a number of purposes. But if you are aiming to build a career in machine learning & data science, you should rather spend more time on mastering Python. Besides, most of the deep learning frameworks you will use such as TensorFlow, PyTorch, and fast.ai are all based in Python.

5. A Portfolio Is a Must

Now, being someone who is trying to enter the data science and machine learning world, this part can be a little tricky. These are new technologies, so, there is a slight chance that you have experience working on related projects. And employers are also aware of it. But that shouldn’t stop you from building a portfolio.

Online portals like Github and Kaggle offer you the platform to showcase your work on whatever individual projects you are pursuing – as a part of developing a new skill. Pretty much, every employer would ask you for Github and Kaggle profiles in the present scenario. So, be ready with them, instead of excuses.

Concluding Remarks

The rapid growth of the digital landscape will continue to require professionals to constantly update their digital skills. For tech-savvy professionals, it means loads of new opportunities, and new horizons of possibilities what they can do with the technology. But in order to take advantage of this dynamism, professionals need to stay abreast with the on-going state of the digital landscape all the times.

Around the world, across the industries, C-suite executives are concerned about the widening digital talent gap in their organization.

AI, big data, cryptocurrency, cyber-security – with so many technologies creating buzz at once, it is becoming increasingly difficult for organizations to determine which skill set they need to invest in.

Here are top 10 essential digital skills that you must have on-board to succeed in 2019, whether you are an organization aiming at transformation or a service provider delivering next-gen services to clients.

Data Analysis

With today’s advanced analytics tools, companies now have the means to analyze heaps of untapped data they have about their customers and organization. But they also need expert data analysts and scientists who can efficiently use these analytics tools to do all sort of analysis work (descriptive, diagnostic, predictive, and prescriptive) on that data, interpret it, and come up with crucial insights.

Besides analysis, the data science team should be especially skilled in the visualization part in order to showcase data and create reports that are easily understandable for the management and can assist them in making decisions.

Data Engineering

Data engineering involves building tools and infrastructure that data analysts/scientists use. While data science focuses on analytics, data engineering is more about data consolidation and warehousing. It is essentially software engineering whose primary purpose is to keep data clean and flowing and deploy data insights at scale.

On the tech side, SQL, Java, Python, Hadoop, and Linux are the hottest data engineering skills currently. In fact, according to a recent study by Stitch Data, the demand for data engineers exceeds the demand of data scientists.

Mobile Expertise

No business needs a reminder that it needs to adopt a mobile-first approach in current dynamics – be it customer apps, content, or internal communication. The mobile computational environment in itself is evolving constantly. So, it is important that developers and marketers stay abreast with the latest mobile trends and be proactive in delivering customers an optimized and state-of-the-art mobile experience.

And in 2019, mobile expertise must not be limited to smartphones or tablets – there is a whole new generation of mobile devices hitting mainstream adoption such as wearables, IoTs, and more.

UX Design

UX design may sound nothing new, but with user’s attention span constantly declining across platforms, focusing on it has become all the more important. In current dynamics, UX isn’t just about visually appealing UI and tried-and-tested navigation. It has become more of a creative-meets-analytical type of role, where every decision is backed by data rather than just guts.

Today’s UX designers also need to think in terms of multi-platform since modern customers’ buying journey span over multiple platforms. So, it is important to deliver a consistent digital customer experience across multiple platforms to ensure a smooth purchase experience.

Machine Learning

Machine learning is unquestionably one of the hottest digital skills in demand today. From voice assistants to data analysis and self-driving cars, there are tons of use cases of this futuristic tech across industries. In fact, all the other digital skills listed here has or may have some use of machine learning for better efficiency.

However, the AI/machine learning ecosystem is quite vast and is mostly exclusive to research currently. Only the supervised learning part of it has corporate applications as of now. So, it is crucial to know the current state of machine learning in the corporate world, how your business can leverage it, only then invest in acquiring the required skilled resources.

Blockchain

Thanks to the Bitcoin buzz, the tech world is now aware of blockchain (even if many still don’t understand it). Blockchain (or distributed ledger) has given rise to decentralized applications, which are inherently more secure and transparent. Apart from its widespread use cases in the finance sector (cryptocurrency), today, the tech community has found a number of other use cases for blockchain such as crowdfunding, file storage, identity management, digital voting, and more.

Building blockchain-based applications requires skills such as networking engineering, cryptography computing, database designing and programming languages (C++, Java, Python, Solidify, etc.).

Related: How AI is driving the next phase of growth in Fintech

AR/VR

AR/VR is already transforming the gaming and entertainment industries and is also gaining wide adoption in media, marketing, advertising, health care, and manufacturing. Businesses of retail, travel, and many other industries have already begun to provide AR capabilities in their apps.

Besides these, AR/VR has opened a whole new world of possibilities how people will consume content in the near future. Currently, video is the most popular mode of content consumption. But as AR/VR based interactive content become easier to create and easier to access, it will naturally surpass video.

Cybersecurity

Data breaches are the biggest threats of the digital age. And when they happen, they often result in long term financial loses for a company. And as the security measures develop and evolve, so do the threats. So, network security or cybersecurity is undoubtedly one of the most important digital skills to have on-board in today’s business environment.

In fact, according to a recent ESG report, cybersecurity has topped the list of the problematic shortage of skills in organizations globally. And over the past few years, the concern has only grown (from 42% in 2015-16 to 53% in 2018-19).

Cloud Computing

Cloud adoption continues to grow. According to LogicMonitor’s Cloud Vision 2020 survey, 83% of enterprise workload will be in the cloud by 2020. To accommodate cloud adoption, migration, and upgrade, organizations need network engineers, cloud architects, developers, and system administrators with relevant cloud computing skills.

However, today the cloud is not the same decade old cloud. From multi-cloud to edge computing, it has evolved a lot and organizations need to keep the latest trend in check and regularly upgrade their cloud strategy to make the most out of it.

Social Selling

Social media has matured over the past decade. It is no longer exclusive to connecting friends and communities. Serious business happens on it every day. Engagement on social media is far better than traditional mediums. For instance, LinkedIn’s InMail open rates are 300% higher than email. And given the nature of the platform, the business world has moved away from hard selling to value-based selling, where mutual trust and relationship with clients/customers is of the highest priority.

So, in current dynamics, having a marketing team with expert social selling skills is must for continuous growth of the organization.

Concluding Remarks

The skill gap is the most prominent threat that looms over the business world today. And multiple reports warn that it will get worse in the near future. It is imperative for C-suite executives to react to this threat and come up with ways to handle the widening digital talent gap in their organization.

And in case we missed mentioning any digital skills that you think are crucial in the current digital age, let us know in the comment section below.

Why are machines so smart?

Because we make them so.

But they can be only as smart as we make them to be.

In today’s AI ecosystem, there is pressure on everyone to make their machine learning algorithms as good as human intelligence. And the only way to achieve it is to have a good amount of quality labeled data to train those algorithms.

Except that the data doesn’t come easy.

Every organization entering into machine learning related services faces this challenge today. And to overcome it, they must have the know-how of different data labeling tools that can help in building quality training data sets and build them efficiently.

Here is a list of the best data labeling tools based on what type of data you are labeling.

Image & Video Labeling Tools

Annotorious

It’s a free, easy to use, MIT-licensed annotation tool for labeling of images on a website. It’s free for commercial use as well. Integrating it with your website only requires adding 2-3 lines of code. You can also explore many of its features in the demos.

Annotorious’ most noticeable features include:

• Image annotation with bounding boxes
• Process maps and high-resolution zoomable images
• Annotorious can be modified with plugins to suit a particular project’s need
• Annotorious Community; where developers can find how they can modify it to extend its capabilities
• Annotorious Selector Pack plugin (to be launched), which will include features like custom shape labels, freehand, point, and Fancy box selection

LabelMe

LabelMe is an open source online data labeling tool. With simple signup, it allows users to label images and share their annotation publically, which is primarily used for a range of computer vision based applications and research.

Some of LabelMe’s key features include:

• LabelMe also offers its mobile app for image labeling & annotation
• Image collection, storage, and labeling
• Training object detectors in real-time
• Simple and intuitive UI
• LabelMe offers MATLAB Toolbox that allows users to download and interact with the images and annotations in the LabelMe database

Labelbox

Labelbox is one of the most versatile labeling tools available today. Its comprehensive features enable organizations to easily adapt and train their machine learning models. Its pricing varies based on the amount of data and the sophistication of the model you are training.

Key features include:

• Labelbox supports Polygon, Rectangle, Line, and Point segmentation, as well as pixel-wise annotation
• You can create bounding boxes and polygons directly on the tiled imagery (zoomable maps)
• Ideal to work with a big team of labelers as it serves up images to be labeled asynchronously, i.e., no two labelers label the same image
• Assured security as the source data is either stored in-house or on a private cloud
• Labelbox allows you to maintain quality standards by keeping track of labeling task performance

Sloth

Sloth is a versatile annotation tool for various data labeling tasks related to computer vision research. It’s free and is one of the most popular tools for facial recognition, therefore, is widely used for surveillance and user identification related applications.

Most notable of Sloth’s features include:

• It allows an unlimited number of labels per image or video frame – leading to more detailed file processing
• It supports various image selection tools – points, rectangles, and polygons
• Developers consider Sloth as a framework and set of standard components that can be configured to build a label tool specifically tailored to one’s needs

Audio Labeling Tools

Praat

Praat is a free audio labeling tool under the Creative Commons (CC BY SA) license, meaning, any derivative works must also come under creative commons license.

Praat’s primary features include:

• Spectral analysis, pitch analysis, format analysis, and intensity analysis of audio files
• It can also identify jitter, shimmer, voice breaks, cochleagram, and excitation pattern
• You can work with sound files of up to 3 hours (2GB)
• It allows you to mark time points in the audio file and annotate these events with text labels in a lightweight and portable TextGrid file
• Users can work with sound and text files at the same time when text annotations are linked with an audio file

Aubio

Aubio is another free and open source annotation tool for audio data labeling. The tool is designed to extract annotations from audio signals. Aubio is written in C and is known to run on most modern architectures and platforms.

Aubio offers the following key features:

• Digital filters, phase vocoder, onset detection, pitch tracking, beat and tempo tracking, mel frequency cepstrum coefficients (MFCC), transient / steady-state separation
• You can segment a sound file before each of its attacks, performing pitch detection, tapping the beat and producing midi streams from live audio
• There’s a dedicated function library to execute above-mentioned functions in real-time applications
• Users can also use these functions offline via sound editors or software samplers

Speechalyzar

Speechalyzar is an audio data labeling tool specifically designed for the daily work of a ‘speech worker’. It can process large speech data sets with respect to transcription, labeling, and annotation. Its main application is the processing of training data for speech recognition and classification models.

Speechalyzar’s main features include:

• You can implement it as a client-server based framework in Java and interfaces software for speech recognition, synthesis, speech classification and quality evaluation
• Speechalyzar also allows you to perform benchmarking tests on speech-to-text, text-to-speech and speech classification software systems
• Ideal for manual processing of large speech datasets

EchoML by Azure and Soundscape are some other audio data labeling tools with rich visualization capabilities that you can also explore.

Suggested: Leverage the power of data with visualization

Text Labeling Tools

Rasa NLU

Rasa NLU is an open-source NLP tool for intent classification and entity extraction. It is primarily used to annotate text for chatbots but can be used for a variety of applications. For instance, recently Trantor used Rasa NLU to train a machine learning model to detect harassment and abuse in email communication within an organization.

Some of the advantages with Rasa NLU are:

• Users can tag multiple words in a single sentence to their respected class or assign the same word in multiple classes
• You can customize and train its language model as per domain-specific needs and get higher accuracy
• Rasa NLU’s open source library runs on premise to keep users’ data safe and secure

Stanford CoreNLP

Stanford CoreNLP is a free, integrated NLP toolkit that provides a set of human language technology tools, which allow users to accomplish various text data pre-processing and analysis tasks.

Here are some advantages with Stanford CoreNLP:

• It offers a broad range of grammatical analysis tools (base forms of words, parts of speech, names, normalize dates, times, and numeric quantities, mark up the structure of sentences in terms of phrases and syntactic dependencies)
• CoreNLP is a fast, robust annotator for arbitrary texts, widely used in production
• It offers a modern, regularly updated package, with the overall highest quality text analytics
• Support for a number of major (human) languages
• It offers APIs for most major modern programming languages and can run as a simple web service

Bella

Bella is a text annotation tool that helps data scientists manage, label, and evaluate natural language datasets. It is designed to save time spent in measuring and learning data, which involves collecting, inspecting, training, and testing data.

Image: A Bella project file for labeling a social media post (Source: Github)

Some plus points of using Bella:

• It offers an intuitive GUI, which allows users to label and tag data through convenient keyboard shortcuts and swipe, and visualize metrics, confusion matrices, and more
• Bella also offers database backend to easily manage labeled data
• Bella is a preferred tool for sentiment analysis, text categorization, entity linking and POS tagging

Tagtog

Tagtog is a versatile text labeling tool that offers manual as well as automated annotation. It’s an AI startup with an impressive client base including AWS, Siemens, and a number of data science research institutions.

Some of the best things about Tagtog are:

• Users don’t require coding knowledge or data engineering concepts to use Tagtog
• Tagtog offers inbuilt ML model to automate text annotation and also provides hassle free deployment and maintenance of manually trained models
• Tagtog annotation tool allows multiple users to collaborate to a single project

Conclusion

There are numerous other data labeling tools in the market, apart from the ones listed above. And as with any other tool for any other purpose, the key is not to know a lot of tools but to know which tool will work best for a given project and to understand how to leverage it best.

And as for which approach you should adopt for labeling – in-house or outsource – that also depends on the project requirements. If you have time and resources, you can do it in-house. If priority is to cater customers with AI driven solutions as quickly as possible, then it is suggested to outsource your projects to a professional firm.

The machine learning market is brewing up and companies are in a rush to get ahead of each other. So, in current dynamics, spending a little to take the advantage of the early bird can make a big difference in the long run.

Need Professional help with your data labeling projects?Let’s Connect!

Today, we have more connected devices than people on the planet.

Even if we exclude our day-to-day connected devices like smartphones, tablets, laptops, etc., there are still more than 7 billion connected IoT devices surrounding us worldwide, deployed in efforts to make our lives more convenient.

In future, we will have even more things connected to the network, including the most common things of every day’s use – water bottles, shoes, hairbrushes, and pretty much anything you can think of.

Cloud’s Incompetence with Connected Devices

These connected devices generate a huge amount of varied and incomplete data that needs to be processed and responded to in a short time. Traditionally, the cloud has been an essential part of this process. But as the need to process IoT data in real-time increased, cloud computing has become a little incompetent for the following reasons:

• Overload: Being centrally deployed on a large scale, cloud platforms usually need to process an enormous amount of data
• Latency: As the physical distance between cloud and device (user) increases, transmission latency increases, so does the response time
• Dependency on the user’s device: With cloud computing, transmission time and processing speed depends highly on the user’s device

Edge Computing to Rescue

Edge computing enables processing of some parts of application & data to be performed by a small ‘edge server’ (referred as FOG Nodes), positioned between the cloud and user (preferably, in a location closer to the user). This allows some of the workload to be offloaded from the cloud as well as from the user’s device – resulting into decreased latency and better performance.

Adoption and Applications

For enterprises aiming at IoT implementation, edge computing has emerged as a great support. Apart from that, various AI applications and emerging 5G communication network also heavily rely on edge computing. As these applications will gain more mainstream adoption, they will also help spur future growth of edge computing.

According to a TrendForce research, edge computing market will grow at CAGR of more than 30% from 2018 to 2022. And this growth is apparent all around us, as edge computing is being used for a wide range of applications. Some examples include:

• Autonomous Vehicles: Autonomous vehicles need to do most of their data processing onboard and in real-time. Depending on the cloud to decide what the vehicle is supposed to be doing in every situation will lead to accidents. Without edge computing, autonomous vehicles are simply not feasible.
• Industrial Automation: In industrial processes, machines generally need to be adjusted as per the surroundings – such as temperature, light, and quality of material coming in. With edge computing capabilities, machines can sense these things and adjust on their own, leading to improved efficiency and extended life of the machine itself.
• Connected Homes/Offices: With Alexa, Google Home, and other voice assistants rapidly becoming a part of our day-to-day life, it is important that their response becomes quicker. Currently, it takes a few seconds for them to respond. With edge computing, their response time will become near real-time.
• Retail: Edge computing is being extensively used in the brick-and-mortar retail space to analyze and respond to customer data in real-time and provide them with a better shopping experience and personalized service.

Recommended: 19 Retail Trends for 2019 [Infographic]

Addressing Security Concerns

A general perception about edge platform’s security is – ‘since data stays in the local environment, (unlike cloud, where it has to travel far through network) security threats go down’. However, on the flipside, edge platform has its own share of security concerns:

a) IoT devices can be compromised by hackers
b) Unlike cloud computing, edge computing data often flows over untrusted public network segments

To address these emerging threats, cybersecurity experts recommend using secure tunnels and VPN to have more control over data in transit. Besides that, using cryptographic keys embedded in IoT device chips for authentication and encrypting local device communication can significantly enhance the data safety within the edge network.

In Conclusion: Overcoming Implementation Challenges

As with any other new technology, edge computing inherently puts new businesses at advantage and imposes threat on the advantages enjoyed by incumbents so far. And their resistance to change only slows down the widespread adoption of the technology, does not stop it.

It is required of businesses to be agile. Adopt edge instead of sticking to the same old practices. Make their processes more efficient, encourage adoption organization-wide, and be proactive to address the security concerns that may emerge with the new ways of doing things.

Planning to switch to edge for your IoT implementation?Let’s talk!

The retail ecosystem is transforming at lightning speed. Advancements in a wide array of technologies are constantly uncovering new possibilities. And with this latest tech easily available to customers, retailers are bound to continuously redefine what ‘good customer experience’ is.

Here is an exhaustive list of retail trends for 2019 and coming years. From new technology to new practices, this list will be a valuable asset to keep handy for all sorts of retailers – be it online, brick-and-mortar, or even mom and pop stores.

Looking forward to upgrading your retail business with these tech trends?Let’s Connect!

Recent advances in computer vision are rapidly taking us towards a future we have long imagined.

A future where you don’t have to feed in or tell computers anything for a range of simple tasks – domestic and industrial. Computers have a vision of their own and can automatically initiate the task you want them to do. [Ex: FaceID by Apple]

Thanks to the advances in deep learning and abundance of visual data available today, computer vision market is one of the fastest growing innovative tech markets today – with a multitude of applications sprouting across multiple industries.

Adopting computer vision is a smart move forward in current dynamics. But don’t just dive in without doing your homework.

To help you with that, here are top computer vision trends that are currently driving value in the sector.

Deep Learning

Deep learning algorithms have multiple advantages over traditional machine learning algorithms. First, they effectively reduce the need for frequent human intervention and thorough domain knowledge in training a model. Second, their workflow allows superior accuracy. And third, the more data you give them, the better results they produce, which is not typical with ML algorithms.

How is it relevant in regards to computer vision — Apart from doing a full picture recognition based on extracted features, they can also identify light and dark pixels, categorize lines and shapes to produce far more accurate results, far more efficiently.

For example, recently, Trantor team did a deep learning project for a leading electric appliance company to automate the extraction of data from electrical blueprints. Our solution provided 95% accuracy and reduced the operation cost for the task by 60%.

Another advantage with deep learning is that the more data (images/videos) you feed in, the better the results get. Many retailers are using these features to map customers’ interaction with products in-store so they can provide them a more personalized shopping experience.

Transfer Learning

Transfer learning has gained a lot of popularity in the recent past. Thanks to its extensive applications in the field of computer vision.

If you are not familiar, it essentially includes using data from one model to train a similar model. The data is fed in the form of layers. These layers are simply properties or constraints related to a specific task.

For instance, if you have a trained ML model A that identifies animals’ pictures, you can use it to train a model D that identifies dogs’ pictures. Or if you want to go a bit complex, then you can use the model D to train a model C that identifies cats’ pictures. In terms of data layers, training D would require adding a few additional layers to A, whereas, training C would require eliminating some dog-specific layers and adding some cat-specific layers to D.

Transfer learning makes things much easier for developers, as you don’t have to work from scratch to train an ML model. And for developers’ convenience, there are thousands of open source ML models, which they can customize to train new models with applications in a number of industries such as retail, healthcare, automobile, transportation, and so on.

Recommended Read: State of Machine Learning (AI) – Notes from DataHack Summit 2018

Point Cloud

Point cloud is a set of data points in 3D space. Simply put, every point on the surface of an object has 3-dimensional coordinates (X, Y, Z), which is referred as point cloud.

Point cloud is a 3D machine vision-based technology, which provides an accurate representation of where an object is in the space. For this reason, it has multiple applications that involve object identification or object movement tracking.

Some of these applications are – monitoring of physical assets, scanning construction sites, scanning landscapes, mapping utility infrastructure, etc. These applications can come quite handy in solving real-world problems like urban planning, repair works, natural disaster management, and so on.

Final Remarks

Visual content provides a firehose of information about the state of the world. And with today’s data technologies, we are able to observe, record, and act upon this information effectively. As a result, the world is changing rapidly, and for any organization, it is crucial to keep up the pace to stay state-of-the-art for itself and its client.

While we have discussed only major developments here, you shouldn’t limit your focus to these alone. We also encourage you to explore areas like mixed reality, edge computing, and semantic segmentation, which are gaining popularity across industries.

Looking to upgrade your business operations with Computer Vision?Let’s Discuss!

Artificial Intelligence is a fascinating subject.

It stimulates our imagination like no other on-going technological advancement.

And as a result, it also gives rise to many speculations. Most interesting (or rather dreadful) of them – “AI will take over the world!”. A notion, which is mostly incepted by Hollywood movies, rather than reality.

Ex Machina movie, in which a humanoid robot outsmarts two highly intelligent people

Fortunately, the reality is not that fascinating. It is exciting, of course, with all the possibilities and on-going developments. But for now, it is mostly an aid to us humans, helping us work more in less time and with better accuracy.

Recently, Trantor’s AI/ML team was at the DataHack Summit 2018 in Bangalore, where many world-class AI experts, data scientists, machine learning engineers, and technology evangelists gathered to discuss ideas, their feasibilities, applications, and much more.

Overall, the event provided many insights on what is relevant in the field of AI today. So, we decided to create this blog to help our clients and tech enthusiasts understand how to make the best use of this technology in current dynamics.

Here are our key takeaways from the event.

Interpretability of Machine Learning Models

Unlike traditional software, the functional part of machine learning algorithms happens in a black box. Meaning, with software, you know exactly what steps led to a specific outcome. But with an ML model, you only have a vague idea of how a particular outcome is achieved.

Since there is no telling exactly how an ML model produced a specific outcome, there is no assurance that it will always outperform a traditional software. And this has been a major hurdle in the adoption of AI/ML so far. For tech guys, it is really difficult to convince management to replace their legacy software with something even they are not sure how does it work.

This is how the scenario looks like in a real organization:

At the DataHack Summit, one of the speakers, Karthikeyan Sankaran, pointed out 3 major reasons for the gap that exists between ML predictions and business decision making. He also suggested solutions that can help bridge this gap; at least up to an extent that allows organizations to leverage ML without much concern. Following diagram highlights these points.

Automated Machine Learning (AutoML)

The workflow of machine learning models involves a lot of steps, such as data pre-processing, feature engineering, feature extraction, and so on. A lot of these steps require an ML expert. And even after these workflow steps, the ML expert has to perform algorithm selection and hyper-parameter optimization at every stage.

And the entire process of ML model training is re-iterative. And it will continue running in search of the most optimum solution for a particular given problem. Clearly, a typical ML workflow is a tedious process. Consequently, it also consumes a lot of time of your skilled employees.

To address this issue, AutoML was introduced. It builds ML models that don’t require human intervention at every stage. It automates the end-to-end process of applying machine learning and frees your skilled resources, produces models faster, which often outperform models designed with the traditional approach. Overall, a win-win solution in pretty much every scenario.

Related Read: How AI Is Driving Next Phase of Growth in Fintech

Transfer Learning

According to Wikipedia, “transfer learning is a research problem in machine learning that focuses on storing knowledge gained while solving one problem and applying it to a different but related problem.”

Simply put, it involves customizing layers of an already trained ML model to solve a new but similar problem. Layers are simply properties or constraints related to that specific problem.

For example, you have a trained ML model A that recognizes dogs’ pictures. Now, if you want another model B to identify wolves’ pictures, you can simple customize certain layers of model A to build model B, instead of building it from the scratch.

Currently, transfer learning is being extensively used in the field of Computer Vision (pertaining to the example above). But its use cases are present in a wide range of AI instances.

In 2018, tools like BERT (Google), ELMO, and ULM Fit were introduced that made transfer learning possible with NLP (Natural Language Processing) models. For example, online retailers can use open source ecommerce chatbot APIs to build chatbots for their own stores. And there are a lot of other open source trained models for developers’ convenience, which they can use to build and train their own models with much ease and in significantly less time – all thanks to transfer learning.

Final Notes & Remarks

Another key insight that we came across at the summit was related to the current industrial state of the unsupervised learning. AI architectures like RBMs (Restricted Boltzmann Machines) and GANs (Generative Adversarial Networks) have created quite a buzz in past, but when it comes to industrial application, they are pretty much in the research stage and mostly exclusive to tech giants.

In fact, unsupervised learning is the part that actually imposes the threat that is often blamed on AI –overpowering human race, taking our jobs, etc. But that scenario is still far – maybe a few decades at least. Yet it may lead to it if we are not careful, as many tech leaders like Elon Musk have warned us about. So, while it is important for the next stage of technology development that unsupervised learning architectures continue to get better, it is also necessary that these developments are regulated. What do you say?

Looking for a trusted technology partner for your AI initiatives?Let’s Connect!

Here is the graphic representation of our recent post on how to get your CRM implementation right.

For further explanation on dos and don’ts mentioned in below infographic, refer to the blog post.