Category Archives for "Blog"

Two women walking down a street with coffee in their hands

What Can US Retailers Learn From Asian E-Commerce Companies?

What Can US Retailers Learn From Asian E-Commerce Companies?

There is stiff competition between Asian e-tailers and their American counterparts for the battle of global leadership. This is an extremely probable future as the e-commerce giants Alibaba and Amazon are already gaining momentum in India and Australia.

Nearly two-thirds of the US e-commerce industry is dominated by retailers such as Walmart, eBay, Best Buy and Amazon (that covers this space by over 40%), the local market is becoming more consolidated and less flexible. This is in comparison to the agile, data-driven, and fragmented Asian e-commerce industry.

Can the mature American retail market learn something from the comparatively young and rapidly growing Asian e-commerce market?

Modify selling strategies to Local markets

US-based brick and mortar stores and online retailers focus on domestic and English-speaking homogeneous markets. On the contrary, the nascent Asian e-tail market is constantly expanding beyond borders to markets with varying population sizes, purchasing powers and cultural backgrounds.

Here’s something to ponder upon — Amazon is the world’s third-largest retailer, it uses its universal selling strategy regardless of the market it scales up to, while its Asian competitor, Alibaba, the world’s sixth-largest retailer, acts on a different vision — “Act Local, Think Global”. This strategy works well in the rather fragmented Asian market and therefore, by extension, in the global market. Alibaba acts through local players or players that know the local market by offering a variety of affiliate programs.

For instance, Alibaba acquired a majority stake in Lazada, a major player in the South East Asian marketplace, an ideal platform for introducing Chinese vendors to non-Chinese buyers. Furthermore, another great example of Alibaba’s adjusting to local markets would be the expansion of Chinese marketplaces Taobao and Tmall in Russia. Alibaba’s strategy envisions adaptations to various local markets and finds ways of making the local systems reinforce each other.

person using smartphone

Cross-Border selling is the name of the game

Alibaba is not the only e-commerce company that wants to increase cross-border sales. Recently, South Korean retailer GS Retail made a $29 million equity investment in the U.S. e-tailer Thrive Market. US e-tailers are not far behind. Even in the US, retailers have begun to realize that buyers outside the English-speaking markets can also generate revenue. This trend could be seen when Walmart, one of the biggest retail chains in the US, acquired a majority stake in India’s largest online retailer, Flipkart, making the transaction “the world’s biggest purchase of an e-commerce company.”

To function in an unfamiliar environment, retailers need expertise, which in today’s market, only the Asian players have. They can soon be well equipped to battle American retail in the global market with optimized operations and the ability to cover different markets with subsidized prices.

Retailers who are tech-savvy, stay ahead

Innovation is a crucial element for businesses to compete in the market whether it wants to play internally, externally or globally. More and more businesses are opting for technology including AI and machine learning to gain an edge over the competition.

ML and AI are disrupting retail by enabling businesses to observe competitive prices and monitoring trends, helping them to react to changes and forecast demand and sales. This way, retailers can boost their revenue and can build data-driven strategies and make better business decisions. No algorithm can be useful if the data it processes is not of high quality. Trained on the data, can it recommend optimal pricing and forecast sales which directly affects the business performance.
The better the data is, the better the outcomes are.

Conclusion

Successful businesses will be those that recognize and adjust their strategies and offerings to that particular market. Moreover, building several channels of communication with customers and leveraging the marketplace as a way of accessing consumers as well as integrating innovations and data into their operations will further strengthen their success.

It’s a no-brainer — Data-driven companies are already dominating the market. The other retailers need to jump on the bandwagon if they want to stay competitive.


Related e-commerce articles

Importance of AI in customer loyalty
The expectations of consumers today from their favourite brands has increased ten-fold, and with big companies such as Amazon providing[...]
What is GS1 Verified?
​Live from the GS1 Connect 2019 EventWith the GS1 Connect 2019 event happening this week in Denver, Colorado, we think[...]
DataPorts and Why they Matter
​Retailers today struggle with managing all of the product content necessary to publish and maintain their online product catalog. Assembling[...]
3 ways Grocery retailers can survive in the age of Amazon
3 ways retailers can survive in the age of AmazonRetail giants like Amazon’s ability to effectively address the ever-changing customer[...]
The Impact of AI on Grocery Retail
The Impact of AI on Grocery RetailIn today’s age, grocery retailers no longer have to make guesses about what customers[...]
Digital Merchandising for Grocery Retail: How can it help boost sales
Leveraging digital merchandising to elevate grocery retailMerchandising is a core skill for both online and offline retailers. Merchandising is the[...]
smiling woman in black and white checker print shirt riding on push cart during daytime

Customer Loyalty Programs: Why Retailers Need Them

Customer Loyalty Programs: Why Retailers Need Them

We have already established that there is cutthroat competition present in the e-commerce and retail industries. This is forcing many brands to re-evaluate how they deliver value and stay relevant to customers. Retailers want to understand what drives their customers to visit their stores and make purchases, and how to reinforce those loyal behaviors.

Loyalty programs is one method to help achieve these goals like increased foot traffic, repeat visitors, build deeper engagement and reap a successful financial return on this loyalty investment. Customer loyalty programs are becoming increasingly popular and they offer a lot of benefits, to both the retailer as well as the customer.

Nielsen found that 84% of consumers are more likely to choose retailers that offer such a program, and 59% report that they’re available where they already shop.

Retailers need to capitalize on this interest in loyalty programs and create an active user base that eventually keeps them engaged with the brand for a long time.

What is a Customer Loyalty Program?

The idea is simple: The more you shop and spend, the more you receive in return. Nielsen describes them as “marketing programs that reward members with purchase incentives.”
With these programs, retailers can track and incentivize purchasing behavior and reward customers for their loyalty to their brand. This is a powerful customer retention tool as it motivates existing customers to continue engaging with the brand and therefore, spend more.

Different types of Customer Loyalty programs

Customer Loyalty programs come in different forms. Some retailers use only one type while some others create combinations of two.

Loyalty Points
This tactic can be seen especially in the grocery chains where customers get points for making purchases or perform certain actions such as providing some personal information.

Social Media
In this approach, retailers abandon the traditional approach to purchasing a product. They award certain points to their customers for social engagement with their brands such as sharing, liking or commenting on an ongoing campaign. Many brands even run contests and raffles that reward loyal fans with amazing prizes.

Paid Programs
Not every reward program is free. Some of these programs require their customers to pay a certain fee that could be a one-time payment or a recurring payment in order to enroll. Amazon Prime is a great example of this type of model.
Furthermore, these programs can also include partnerships with credit card companies who may provide special benefits and offers in exchange for reward points. Some of these benefits may include discounts, cash-backs, free shipping, access to exclusive shopping events, free services, upgrades.

Retailers may use these programs to modify buying behavior by incentivizing the action they want their customers to take. These programs also provide data to help retailers understand their customers more deeply. With this kind of data on purchasing behavior, it’s easier to segment, create personas, and gain insights to help guide new initiatives.

Role of AI in Customer Loyalty Programs

AI has found its way to many retail companies across different verticals and now have slowly made their way into loyalty and marketing programs as well.

Customers as well, to an extent have become familiar and comfortable with using these technologies. A research states that customers are increasingly willing to rely on algorithms and smart devices for enhanced and personalized retail experiences. This, in turn, fosters an expectation for convenient, low-friction shopping experiences with loyalty programs. AI and machine learning may help in streamlining customer experiences, but they are apt for managing and interpreting customer data captured by loyalty activities and customer interactions. A marketing strategy with an integrated AI and machine learning technology can create a single customer view dynamically, in real time. This can help brands and retailers with large footprints or multiple locations can understand their customers faster and predict trends and offers accordingly. Moreover, this helps them to stay ahead of the competition. Of course, enthusiasm for these technologies is at a high point, and there are many varied predictions about the impact AI will have on the world at large.

Summary

A research shows that retailers spend 5 to 10 times more to capture a new customer than to retain the currents ones. With Customer loyalty programs, engaging with the existing customers could cost less, and reap larger benefits in the future. Effective execution of these programs can increase the customer lifetime value and ROI. There is a huge chunk of consumers that modify their spending amounts in order to maximize points. Hence, program members are more likely to shop on a regular basis. Furthermore, they also activate word of mouth marketing as one customer’s experience with a brand can influence another’s choice or preference for a brand.

Related Articles

Enhancing the customer experience beyond shopping
Enhancing the customer experience beyond shopping With consumers changing how they purchase and engage with a brand, retailers have leveraged[...]
Importance of AI in customer loyalty
The expectations of consumers today from their favourite brands has increased ten-fold, and with big companies such as Amazon providing[...]
What is GS1 Verified?
​Live from the GS1 Connect 2019 EventWith the GS1 Connect 2019 event happening this week in Denver, Colorado, we think[...]
DataPorts and Why they Matter
​Retailers today struggle with managing all of the product content necessary to publish and maintain their online product catalog. Assembling[...]
3 ways Grocery retailers can survive in the age of Amazon
3 ways retailers can survive in the age of AmazonRetail giants like Amazon’s ability to effectively address the ever-changing customer[...]
The Impact of AI on Grocery Retail
The Impact of AI on Grocery RetailIn today’s age, grocery retailers no longer have to make guesses about what customers[...]
woman looking at phone and smiling

How important is catalog quality in e-commerce?

How important is catalog quality in e-commerce?

Lead nurturing is one of the most important aspects of E-commerce. A lot of retail brands so far have been investing their focus on lead acquisition investing in various sources via paid or organic means to lead potential customers to their landing pages. It is imperative for brands to pay attention to nurturing the lead until a conversion takes place. Here is when catalog quality plays an important role.

Catalog quality impacts the customer journey

A vital part of the process involves the catalog present on the e-commerce website. A weak catalog can cost customers and lose potential leads.  The customer journey from the time they land on the website, to the final checkout depends heavily on the catalog quality of the website, as it guides the customer through each step of the sales process. It is important that the customer does not leave at any stage of the conversion funnel. From the product list page to the payment page - lead nurturing can happen by an improved catalog quality.  

Some of the things to keep in mind - products must be properly classified on the landing page, product attributes must be well defined and images must be well represented. Furthermore, the information provided on the product page should be clear and unambiguous.

woman sitting on bed facing laptop
Content matters for catalog optimization

Catalog monitoring helps in tracking the catalog health ensuring product classification, data completeness, avoiding duplication and identifying errors in product information. Catalog optimization can help personalize a customer’s journey through the conversion funnel. This can be done by customizing product titles as per the customer’s search keywords as well as their descriptions, replacing missing titles and descriptions, enriching product information, and tags.

Catalogs are the focal point when a customer is making a buying decision. They take into consideration all the available details while choosing a product. If there is any discrepancy in the details, such as a missing feature can cost a potential lead, as the customer may abandon his journey.

Moreover, maintaining consistent catalog health can help improve conversion rates, by not missing out on qualified leads. Around 70% leads coming through any channel may not be all qualified leads, in spite of the credibility of its sources. Companies who invest in nurturing their leads can see a 450% increase in qualified leads and a 50% increase in sales.

Ultimately, a strong catalog quality can define a customer's experience on your site. This is what can impact a customer's revisit to your website. Building a good catalog helps towards building a strong relationship with a customer.

 

With CatalogIQ, we can help your e-commerce or retail business by strengthening your content using AI and machine learning.
Are you looking for boosting your eCommerce business? At Icecream Labs, we leverage AI and machine learning to provide solutions for your business problems. To know more on how AI can impact your business you can get in touch on LinkedIn, or mail us — sales@icecreamlabs.com

Related e-commerce articles

Enhancing the customer experience beyond shopping
Enhancing the customer experience beyond shopping With consumers changing how they purchase and engage with a brand, retailers have leveraged[...]
Importance of AI in customer loyalty
The expectations of consumers today from their favourite brands has increased ten-fold, and with big companies such as Amazon providing[...]
What is GS1 Verified?
​Live from the GS1 Connect 2019 EventWith the GS1 Connect 2019 event happening this week in Denver, Colorado, we think[...]
DataPorts and Why they Matter
​Retailers today struggle with managing all of the product content necessary to publish and maintain their online product catalog. Assembling[...]
3 ways Grocery retailers can survive in the age of Amazon
3 ways retailers can survive in the age of AmazonRetail giants like Amazon’s ability to effectively address the ever-changing customer[...]
The Impact of AI on Grocery Retail
The Impact of AI on Grocery RetailIn today’s age, grocery retailers no longer have to make guesses about what customers[...]
clothes rack with a blurred background

Data Quality – How to measure for best results?

Data Quality - How to measure for best results?

There is a constant flow of data coming in to be extracted, and decisions have to be made. Ensuring that the data is accurate can become cumbersome if certain rules are not put in place. Data quality is a delicate balancing act juggling between accuracy and completeness. 

From an e-commerce perspective, Search Engine Algorithms need an understanding of the context behind each of the search terms that would inadvertently lead to better results and bag conversion rates. Attributes that are provided by the product content help search engines to understand the context as well as the intent of a consumer behind a  specific search term.

Many retail organizations manually add structured attributes using BPO companies or crowdsourcing firms like Amazon Mechanical Turk. Some large retail corporations engage their own associates to manually validate the quality of the data provided by their suppliers before the products are made available for customers.

How to effectively assess the data quality of product content?

A comprehensive yet concise data quality checklist helps in working towards analyzing the data. DAMA UK created an excellent guide on “data dimensions” that can be used to get a better picture of how data quality is determined.

There are 6 dimensions or steps that data quality can be determined from - 

  • Completeness - If the data coverage across required fields available
  • Uniqueness - When measured against other data sets, there is only one entry of its kind.
  • Accuracy - How well does the data reflect the real-world person or thing that is identified by it? 
  • check
    Timeliness - This could be previous sales, product launches or any information that is relied on over a period of time to be accurate.
  • check
    Validity - Does the data conform to the respective standards set for it?
  • check
    Consistency - How well does the data align with a preconceived pattern? 

In this article, we will measure data against 3 of these dimensions - 

Accuracy:

Improving data to produce the right information is the need of the hour. Accuracy states that the data is what it should be or it is a percentage measurement of how accurate the data is.

On using the search query - “floral cardigan” across 2-3 e-commerce sites, the search terms were classified individually wherein floral was classified as a pattern and cardigan as the clothing. This lead to incorrect results on the first page itself.

web page results for cardigans

Enhancing accuracy is a benefit to all, having more accurate attributes, as well as accurate values, enables search engines to produce better results. This eventually impacts and enhances the customer experience as well as bag conversions.

Completeness:

Completeness is the measure of whether data searched for exists or does not. It drives the attribute definition across all products as well as the coverage of attribute values across all available fields present in the product content.

For instance, if we have an attribute “pattern” for all products under product type skirts that contains 100k products, but pattern as an attribute is added only for 60k products, then “pattern” is only 60% filled in.

If we now perform a search for “striped cocktail dresses” at Macy’s, we could expect would look something like this  -

web page, cocktail dresses

However, these are the results that you get instead - 

web page, cocktail dresses

Investing efforts in encapsulating key data sets as well as attribute values while loading products from product content from suppliers would reap rewarding results. The completeness of key attributes and their values is very important. Missing data can cost a lot of potential customers.  

Consistency:

Consistency refers to data that requires to follow the same format for all the attributes in the product content which needs to match internal and external standards. Maintaining standard data formats across product types as well as attributes and values; keeping it in line with attribute names as well as the value label in accordance with external standards would help to show customers their desired products easily, without friction.

For instance, searching for “fruity perfumes for women” on Target results with these products and does not change even if another search query is given.  This gives the impression that the different scent attributes have not been cataloged.  Furthermore, the search query provides the same number of results despite the different search queries.

web page for perfumes
web page for perfumes

Values for scents provided on the product page are inconsistent and do not conform to external standards; although the “scents” may not be indexed for identification by the search engine.

There are certain attributes and values that would require consistency, which may be measured for reliability reasons. Measuring conformity against internal standards can be enhanced by ensuring the adherence to adding only valid labels that are verified and consistent with external standards. This would help to refer to the same product content available across retailers and define the list of standard attributes and their labels.


Conclusion

In conclusion, as more and more products along with the categories are being loaded onto platforms, attributes and their respective values are critical to understanding the customer’s purchase intent.

It is also essential to assess catalog data quality regularly so the three aspects of - consistency, completeness and accuracy - which help customers get better results and e-commerce portals can efficiently increase conversion rates.


Related Articles

Enhancing the customer experience beyond shopping
Enhancing the customer experience beyond shopping With consumers changing how they purchase and engage with a brand, retailers have leveraged[...]
Importance of AI in customer loyalty
The expectations of consumers today from their favourite brands has increased ten-fold, and with big companies such as Amazon providing[...]
What is GS1 Verified?
​Live from the GS1 Connect 2019 EventWith the GS1 Connect 2019 event happening this week in Denver, Colorado, we think[...]
DataPorts and Why they Matter
​Retailers today struggle with managing all of the product content necessary to publish and maintain their online product catalog. Assembling[...]
3 ways Grocery retailers can survive in the age of Amazon
3 ways retailers can survive in the age of AmazonRetail giants like Amazon’s ability to effectively address the ever-changing customer[...]
The Impact of AI on Grocery Retail
The Impact of AI on Grocery RetailIn today’s age, grocery retailers no longer have to make guesses about what customers[...]
gift box

Disrupting the gifting industry with G-commerce

Disrupting the gifting industry with G-commerce

Gifting isn’t the same as it used to be. Thanks to last minute gifting. This has led to a massive change in the gifting industry, with Loop Commerce.

Loop Commerce is disrupting the $200 — $300 billion gifting market with their understanding of gifting commerce — changing the gift buying and receiving experience for customers. Loop Commerce have reasons to believe that all consumers have different mindsets while they shop and so the experience can’t be the same.

lady holding tablet, shopping on app

The platform takes care of virtual gifting by making all products easily giftable. Gift recipients can unbox their gifts virtually through a beautiful experience online. They can also exchange their gift with something else. One of the main issues retailers face is a decrease in the number of transactions per customer that results in lesser sales. Professor Dan Ariely, expert in Behavioral Economics works closely with Loopcommerce, a lot of studies on consumer psychology and behavioral economics confirm that gifting is an emotional process.

Loop Commerce data suggests that around 40% of the gifts are purchased on the day of the occasion or the day before, since a lot of people forget to buy gifts. Many retailers don’t take this last minute gifting habit into consideration , which leads to customer drop off. Some of the largest retailers are slowly adopting Loop’s solution. Implementing G-Commerce can help retailers increase customer satisfaction and double conversions.

Related Articles

Enhancing the customer experience beyond shopping
Enhancing the customer experience beyond shopping With consumers changing how they purchase and engage with a brand, retailers have leveraged[...]
Importance of AI in customer loyalty
The expectations of consumers today from their favourite brands has increased ten-fold, and with big companies such as Amazon providing[...]
What is GS1 Verified?
​Live from the GS1 Connect 2019 EventWith the GS1 Connect 2019 event happening this week in Denver, Colorado, we think[...]
DataPorts and Why they Matter
​Retailers today struggle with managing all of the product content necessary to publish and maintain their online product catalog. Assembling[...]
3 ways Grocery retailers can survive in the age of Amazon
3 ways retailers can survive in the age of AmazonRetail giants like Amazon’s ability to effectively address the ever-changing customer[...]
The Impact of AI on Grocery Retail
The Impact of AI on Grocery RetailIn today’s age, grocery retailers no longer have to make guesses about what customers[...]
code, matrix, laptop

Application of Artificial Intelligence to Extract Web Data

Application of Artificial Intelligence to Extract Web Data

The web is a giant storehouse where data is available in abundance. As on 2018, there are nearly 5 billion websites, according to a World Wide Web survey. The potential of what can be done with this data is massive but it is practically impossible to access this data manually. The main challenge that is being faced right now is navigating through this unstructured pile of information and extracting it. This takes a lot of time and effort especially scraping the data from the web. Here is the time when automation can play a pivotal role.

With the rapidly changing technological trends and advancements in AI, there is a way to use machines to extract information from a variety of sources on the web and train them to do it on their own.

Here is a quick example to make this simpler to understand — when we scan and skim through a document for a specific piece of information, we additionally look for alternative sources as well. This inadvertently adds to our knowledge on the topic. The AI system works in just that manner.

Automation is important for web data crawling

A major advantage of AI powered web crawlers are that they save cost of manpower and time put in manually. This even reduces the probability of human error. For simple web data crawling, it is a given that softwares can do this task faster and more accurately than humans. Once a data crawler machine is trained, it can efficiently extract from every single source.

This machine should be able to navigate through different pages and should collect the data from each of them. This is where things may get difficult as different websites use different navigation systems which results in complexity, hence the programmer, writing this code must have sound technical knowledge. He must deploy the code in such a manner that there should be minimal human interference once the machine is programmed.

The future of data extraction

With the ever growing need for data and the challenges associated with procuring it, AI can be the missing piece of the puzzle. The research behind this has tremendous potential with a positive glimpse into the future where intelligent machines with human sight can crawl web documents to give the missing pieces of information that we need to know.

The AI system can be a game changer in research tasks that require a lot of human labor. A system like this will not only reduce the time taken but also enable us to use the abundant information on the web.

Looking ahead, this new research is only a step towards creating the truly intelligent web crawler that would eventually be able to master a variety of tasks just like humans rather than being focused at just one process.

The application of AI and automation in web data crawling and scraping is extensive. Compared to humans, the consistency of the AI powered web data crawlers is unmatched. Furthermore, these machines do not require a lot of maintenance over long periods of time, which adds to its value. There is an immense amount of potential for improvements in web data crawling automation leveraging AI and therefore, the possibilities are endless.

Related AI articles:

Enhancing the customer experience beyond shopping
Enhancing the customer experience beyond shopping With consumers changing how they purchase and engage with a brand, retailers have leveraged[...]
Importance of AI in customer loyalty
The expectations of consumers today from their favourite brands has increased ten-fold, and with big companies such as Amazon providing[...]
What is GS1 Verified?
​Live from the GS1 Connect 2019 EventWith the GS1 Connect 2019 event happening this week in Denver, Colorado, we think[...]
DataPorts and Why they Matter
​Retailers today struggle with managing all of the product content necessary to publish and maintain their online product catalog. Assembling[...]
3 ways Grocery retailers can survive in the age of Amazon
3 ways retailers can survive in the age of AmazonRetail giants like Amazon’s ability to effectively address the ever-changing customer[...]
The Impact of AI on Grocery Retail
The Impact of AI on Grocery RetailIn today’s age, grocery retailers no longer have to make guesses about what customers[...]
shopping cart in a grocery store

Blockchain — The Retail Advantage

Blockchain — The Retail Advantage

The retail landscape has changed over the last decade. With newer technological enhancements, more retailers are opting to incorporate the latest technology to stay ahead of the competition.

From drone deliveries to one hour deliveries, companies are increasingly investing in virtual assistants and AI to engage with customers and enhancing customer experience.

Blockchain is not far behind. With its satisfying results by effectively creating a complete virtual financial market, it is here to stay. It’s no surprise that blockchain is revolutionary technology. The retail industry has finally recognised the power blockchain holds.

Here’s how blockchain can play a crucial role in the future of retail:

Data Collection and Analysis

Data enhances the shopping experience for customers. Blockchain does that, by delivering an efficient way to collect and analyse the available information. Leveraging AI, Blockchain can gather and assess data in real time from different sources like consumers and retailers.

There is a lot of data now that is available from different locations. Unfortunately, the data that is available is fragmented. This makes it difficult to sift through and detect patterns that can let the retailer know which direction they have to take to enhance their customers’ experience and address their pain points. These become missed opportunities for retailers. Here, the blockchain technology can aptly address this problem and make it easier. A blockchain platform collects data from across the supply chain and leverages machine learning to structure the data. Blockchain can enhance the inventory tracking process, including reducing overstocking and under-stocking. Since blockchain uses a secure ledger format, the product data is more reliable and secure from tampering. Furthermore, it can reduce supply chain product waste.

blockchain-technology

Supply Chain Ledger

There are several ways that blockchain can help strengthen relationships. The entire supply chain is one major aspect retailers need to keep secure as it directly impacts the shopping experience for the customers. Using blockchain as a supply chain ledger can make a huge difference in all segments of retail. This especially can impact and enhance retailers that are involved in perishable goods.

Blockchain ensures that the supply chain and logistics is secure and authentic. This means that every record and form is being checked and time stamped in the supply chain. This ensures no tampering of data with everything being independently verified. Thus, there is greater control over product manufacturing location, process, and timing.

Furthermore, a blockchain supply chain model also enables retailers to control all aspects of transportation, storage, delivery, and presentation.

Payments and e-commerce

Blockchain is a trusted means of payments. The majority of retailers are integrating bitcoin and other cryptocurrencies as means of payment processing. The big advantage here is, compared to credit cards, the integration of cryptocurrencies is that it is cheaper and transparent to process transactions.

Blockchain allows for retailers to accept cryptocurrencies along with digital records which helps streamline refunds and return processes.

Besides, purchasing items that need a large amount of money such as cars or land property with cryptocurrency can track the ownership and verify resale of stolen goods.

e-commerce

Retailers must now realise that Blockchain is here to stay. While there are other developments happening in the industry, they need to keep an eye out for this technology.

The possibilities have only just begun.

Related AI Articles:

convolution filter or matrix

3×3 convolution filters - A popular choice

In image processing, a kernel, convolution matrix, or mask is a small matrix. It is used for blurring, sharpening, embossing, edge detection, and more. This is accomplished by doing a convolution between a kernel and an image.

In this article, here are some conventions that we are following —

  • We are specifically referring to 2D convolutions that are usually applied on 2 matrix objects such as images. These concepts also apply for 1D and 3D convolutions, but may not correlate directly.
  • While applying 2D convolutions like 3X3 convolutions on images, a 3X3 convolution filter, in general will always have a third dimension in size. This filter depends on (and is equal to) the number of channels of the input image. So, we apply a 3X3X1 convolution filter on gray-scale images (the number of channels = 1) whereas, we apply a 3X3X3 convolution filter on a colored image (the number of channels = 3).
  • We will refer to all the convolutions by their first two dimensions, irrespective of the channels. (We are observing the assumption of zero padding).

A convolution filter passes over all the pixels of the image in such a manner that, at a given time, we take 'dot product' of the convolution filter and the image pixels to get one final value output. We do this hoping that the weights (or values) in the convolution filter, when multiplied with corresponding image pixels, gives us a value that best represents those image pixels. We can think of each convolution filter as extracting some kind of feature from the image.

Therefore, convolutions are done usually keeping these two things in mind -

  • Most of the features in an image are usually local. Therefore, it makes sense to take few local pixels at once and apply convolutions.
  • Most of the features may be found in more than one place in an image. This means that it makes sense to use a single kernel all over the image, hoping to extract that feature in different parts of the image.

Now that we have convolution filter sizes as one of the hyper-parameters to choose from, the choice is can be made between smaller or larger filter size.

Here are the things to consider while choosing the size —

Smaller Filter Sizes

Larger Filter Sizes

We only look at very few pixels at a time. Therefore, there is a smaller receptive field per layer.

We look at lot of pixels at a time. Therefore, there is a larger receptive field per layer.

The features that would be extracted will be highly local and may not have a more general overview of the image. This helps capture smaller, complex features in the image. 

The features that would be extracted will be generic, and spread across the image. This helps capture the basic components in the image. 

The amount of information or features extracted will be vast, which can be further useful in later layers.

The amount of information or features extracted are considerably lesser (as the dimension of the next layer reduces greatly) and the amount of features we procure is greater.

In an extreme scenario, using a 1x1 convolution is like considering that each pixel can give us some useful feature independently. 

In an extreme scenario, if we use a convolution filter equal to the size of the image, we will have essentially converted a convolution to a fully connected layer. 

Here, we have better weight sharing, thanks to the smaller convolution size that is applied on the complete image. 

Here, we have poor weight sharing, due to the larger convolution size.

Now that you have a general idea about the extraction using different sizes, we will follow this up with an experiment convolution of 3X3 and 5X5 —

Smaller Filter Sizes

Larger Filter Sizes

If we apply 3x3 kernel twice to get a final value, we actually used (3x3 + 3x3) weights. So, with smaller kernel sizes, we get lower number of weights and more number of layers. 

If we apply 5x5 kernel once, we actually used 25 (5x5) weights. So, with larger kernel sizes, we get a higher number of weights but lower number of layers. 

Due to the lower number of weights, this is computationally efficient. 

Due to the higher number of weights, this is computationally expensive. 

Due to the larger number of layers, it learns complex, more non-linear features.

Due to the lower number of layers, it learns simpler non linear features.`

With more number of layers, it will have to keep each of those layers in the memory to perform backpropogation. This necessitates the need for larger storage memory. 

With lower number of layers, it will use less storage memory for backpropogation.

Based on the points listed in the above table and from experimentation, smaller kernel filter sizes are a popular choice over larger sizes.

Another question could be the preference for odd number filters or kernels over 2X2 or 4X4.

The explanation for that is that though we may use even sized filters, odd filters are preferable because if we were to consider the final output pixel (of next layer) that was obtained by convolving on the previous layer pixels, all the previous layer pixels would be symmetrically around the output pixel. Without this symmetry, we will have to account for distortions across the layers. This will happen due to the usage of an even sized kernel. Therefore, even sized kernel filters aren’t preferred.

1X1 is eliminated from the list as the features extracted from it would be fine grained and local, with no consideration for the neighboring pixels. Hence, 3X3 works in most cases, and it often is the popular choice.

Related AI articles

Enhancing the customer experience beyond shopping
Enhancing the customer experience beyond shopping With consumers changing how they purchase and engage with a brand, retailers have leveraged[...]
Importance of AI in customer loyalty
The expectations of consumers today from their favourite brands has increased ten-fold, and with big companies such as Amazon providing[...]
What is GS1 Verified?
​Live from the GS1 Connect 2019 EventWith the GS1 Connect 2019 event happening this week in Denver, Colorado, we think[...]
DataPorts and Why they Matter
​Retailers today struggle with managing all of the product content necessary to publish and maintain their online product catalog. Assembling[...]
3 ways Grocery retailers can survive in the age of Amazon
3 ways retailers can survive in the age of AmazonRetail giants like Amazon’s ability to effectively address the ever-changing customer[...]
The Impact of AI on Grocery Retail
The Impact of AI on Grocery RetailIn today’s age, grocery retailers no longer have to make guesses about what customers[...]
wooden blocks

Capsule Network — Better approach for Deep Learning

Capsule Network — Better approach for Deep Learning

Deep learning is a concept of the machine learning family. It supports the statement “learning can supervised, semi-supervised or even unsupervised rather attaching oneself to the task-specific algorithms. These algorithms are then, used to define the complex relations like that of the human biological nervous network. The interpretation of deep learning is established on the concepts of information processing and communication patterns.

The architecture of Deep learning was introduced in the fields of computers and technology. It was then rapidly adapted and implemented with various areas such as in speech recognition, social network filtering, bioinformatics, drug design, and similar concepts which involve artificial intelligence. The results after implementing Deep learning were remarkable. The machines processed the desired results with better speed and accuracy than humans. There have also been cases where these machines have proven to be better than human experts.

Routing and networking has taken a new direction

Last year, Geoffrey Hinton, one of the Godfathers of deep learning released a paper — Dynamic Routing Between Capsules”. The paper compared to the current state-of-the-art Convolutional Neural Networks, the authors projected that the human brain have modules called “capsules”. These ‘capsules’ are adept at handling different types of visual stimulus like pose (position, size, orientation), deformation, velocity, hue, texture etc. The brain must have a mechanism for “routing” low level visual information to what it believes is the best capsule for handling it.

Convolutional Neural Networks (CNN) vs capsule network

The concept of Capsule network relates to the dynamic routing between capsules unlike in Convolutional Neural Network (CNN) wherein connectivity patterns are derived between neurons or stimuli.

The term capsule refers to a collection of neurons whose activity vector represents the visual field or the instance parameters of the object or the object part.

The key feature of CNN lies in the recommendation systems, image/ video recognition and natural language processing, etc. but with Capsnet, enabled by Deep learning methods designs images in an objectified manner rather than being a sketched version.

There is a strong correlation between the hidden layers of the entity or the distant object in CNN. The visualization of the object references is made better with the Capsule Network.

Capsule is a nested set of neural layers.

In a regular neural network, more layers can be added. In CapsNet the extra layers are added inside a single layer. The neural networks use minimal pre-processing as compared to other image classification algorithms. This states that the networking learning filters used in the traditional algorithms were predominantly hand-engineered.

CNN is what created a major change and proved to be an advantage to the human efforts in feature design. It also made it easier for humans who didn’t have prior knowledge about feature design.

With the introduction of the CapsNet, the task has become much simpler. Capsnet observes a strong correlation among the varied layers of the interface. This correlation takes a new dimension in the image processing techniques. The layers are arranged in functional pods which enable designers to distinguish between the various elements.

For instance, the approximate positions of the nose, mouth, eyes can be drawn using CNN but the exact alignment or the 3D structure can be fixed with CapsNet.

Image by Aurélien Géron

 

Practical Advantages of CapsNet in Deep learning

The probabilistic interpretation is derived from machine learning. Its features include optimizing the training and testing concepts. To be more specific, the probabilistic interpretation considers the non linearity activity as a cumulative distribution function. This concept led to the covering up glitches or dropouts in neural networks.

The capsules try to resolve the problem with max pooling levels through equivariance. This indicates that capsules make it viewpoint-equivariant, instead of making the function translation invariance. So as the function goes and changes its place in the picture, the function vector reflection will also alter in the same way which makes it equivariant.

sketchbook with a sketch of a robot

The core components of Artificial Intelligence

The core components of Artificial Intelligence

The massive surge in AI and the buzz surrounding it sometimes makes it difficult to get a handle on the technology and piece around AI. Here we make a small attempt to explain the core components, services and piece in AI.


Hardware

Graphic Processing Units or GPUs - Nvidia is the largest maker of GPUs had its niche in the gaming market. It had such a lasting impression that no gamer would want to be caught without it. This was until the discovery of Bitcoin and the heavy interest in blockchain. GPUs were no longer restricted to gaming. People relied on them to mine bitcoins.
With the current buzz regarding AI and the rush towards adapting it in their processes, people have realized the benefit of using GPUs for AI and deep learning. GPUs essentially are high end graphic cards that go on to regular servers. The cards as well as the extensive software stack make it easy to process large volumes of data in complex AI models. GPUs are expensive but they cut processing time from months to days. Now, you would never see a deep learning developer without a GPU. Nvidia from its end, has built an extensive software stack to support the use of these GPUs. Libraries such as Cuda are a great resource.

Frameworks

While the hardware is in place, the framework or libraries are the what the machine learning applications are built on. Matlab was widely used to experiment and the R programming was also heavily used.
When Python libraries became available, developers switched almost immediately. Scikit-Learn is the most popular python framework most developers use to do traditional machine learning.
The big push towards deep learning has lead to some great competing frameworks. Most recently, Tensorflow from Google has gained popularity as well as Pytorch. Some others in the same field are Kara’s, Caffe and Theao. Each frameworks have their own set of pros and cons.
Every framework has its pro and cons. The adoption is really driven by the community available for people get support since all of these are free open sourced and not commercially sold frameworks. Tensorflow is pushing the limits with awesome features to handle text, images etc and has been widely used within Google. Our teams here are currently in love with Pytorch.

APIs

All the big cloud providers such as Amazon Web Services (AWS), Google Web Platform or Microsoft Azure have a portfolio of AI or machine learning APIs offered as cloud services. They offer text classification, sentiment analysis, image classification etc. These services can be plugged into solve simple problems within a developers application. For instance, for a broad level search, you can use the sentiment analysis API to see if your customer feedback is negative or positive.

AI Applications

These are the application used by customers which leverage AI and solve specific problems. Siri and Alexa are the ideal examples. Another great example would be the recommendation engines that we see on Amazon, Walmart, etc. The most common applications can be seen in almost every car leveraging technology from MobileEye. It has features such as the lane assist, parking sensor, and pedestrian or obstacle detection systems.
Our teams at IceCream Labs have spent the better part of the last 18 months building applications focused on catalog management and merchandising for retailers and brands.

alexa

We leverage GPUs, Tensorflow, Pytorch, Keras and Caffe. We don’t use the standard cloud APIs as they are not sufficient for the problem we cover. Our application, CatalogIQ can intelligently score the quality of the product content and automatically classify products, generate keywords, improve content for SEO and search.

By leveraging secure private clouds, we can do this on from 1 product to 100 million products seamlessly. That is, really the power of AI.