verbinde dich mit uns

TECHNOLOGIE

How to Detect and Eradicate Bias in Natural Language Processing (NLP)?

Veröffentlicht

An

How to Detect and Eradicate Bias in Natural Language Processing (NLP)?

Natural Language Processing (NLP) is one of the main components of artificial intelligence (AI) and machine learning (ML).

Like those two, NLP-powered systems also reinforce certain biases primarily because of the data used for machine learning. The element of bias in NLP systems can be reduced if organizations and AI experts work in coordination with one another for the same.

Natural Language Processing (NLP) systems are highly useful in our day-to-day lives as well as digitized and automated business operations. NLP, a powerful subfield of AI, computer science and linguistics, can be applied in fields as disparate as supply chain management Und tourism.

Despite its vast number of use cases, NLP, like AI, has tendencies to reinforce and exploit certain biases present in the underlying data used for NLP-based systems. Even powerful NLP tools such as the mighty GPT-3 show this problem. To understand this better, suppose an NLP tool is used as a resume filtering system in a recruitment campaign conducted by an organization. Biased NLP may exclude women and candidates of color, and only select a certain type of individual (white male, in this example) for employment. Recruitment-related biases, in either NLP or otherwise, can be disastrous for organizations.

As we know, biases are mostly prevalent due to a lack of training datasets and, more importantly, in the case of NLP, because social biases are ever-present in the languages we speak and write. As a result, such biases can surreptitiously enter NLP systems too. To detect and reduce the element of bias in NLP systems, the following ideas can be used:

1. Implement Audits to Track Biases

AI experts need to carry out audits to discover biases—and their magnitude—in data generated from NLP systems. For example, such audits can be highly useful for understanding the underlying biases that social media users show in their posts. Additionally, such audits can also allow concerned officials in public or private agencies to know about text related to racism or other marginalization-fueled speeches on a public platform.

LESEN  How bias in AI can damage marketing data and what you can do about it

2. Establish AI Model Training Standards

As stated earlier, biases in NLP are mainly caused by the kinds of data that are used in model training. As a result, organizations need to monitor the datasets that are being used for the purpose of NLP model training. If the bias-related elements are removed from AI or NLP models, then the overall bias in NLP systems will also reduce drastically.

Apart from these, data analysts can also use data security assessments to find if NLP datasets are trained with authentic and valid natural language data and not other, possibly contaminated data. Eliminating bias in NLP systems and models is a seemingly impossible task. However, by adopting intelligent data governance and quality control during the early phases of AI and NLP implementation, it can be reduced up to a great extent.


Quellenlink

Behalten Sie im Auge, was wir tun
Seien Sie der Erste, der neueste Updates und exklusive Inhalte direkt in Ihren E-Mail-Posteingang erhält.
Wir versprechen, Sie nicht zuzuspammen. Sie können sich jederzeit abmelden.
Ungültige E-Mail-Adresse
Zum Kommentieren klicken

Hinterlasse eine Antwort

Ihre E-Mail-Adresse wird nicht veröffentlicht. Erforderliche Felder sind mit * markiert

TECHNOLOGIE

Developer Community Key to Powering Workplace Innovation

Veröffentlicht

An

Developer Community Key to Powering Workplace Innovation

Workday DevCon 2023 showcased the power of partnerships, AI and ML, and community in driving workplace innovation.

My degree in international business and MBA in finance did not teach me how to code.

I was an accountant for a tech company during the early days of my career and that meant learning to program on my own. My desk looked like a reference librarian’s with two-inch-thick product guides and technical manuals perched on my desk. I was lucky. I had great mentors (thank you, David, for patiently teaching me to design top performing applications!) and something else unique to many developers: endless curiosity. 

Thankfully, today is different (except for the curiosity part), in part due to companies like Workday.

Workday champions its developer community much like David invested in me. Recently, hundreds of developers gathered at Workday DevCon 2023 in beautiful San Jose, California. It was mentorship at scale.

As one would expect, Workday broke some big news like the availability of App Builder — Workday’s no-/low-code tool to collaboratively, quickly, and easily build Workday Extend Apps on the web as well as GraphAPI, which leverages GraphQL technology to provide a single pane of glass for developers to discover and access Workday data. 

The event, however, went far beyond the technology stack. Let’s take a look at DevCon through the lens of three critical power centers: partnerships, AI and ML, and community.

The Power of Partnerships

Matthew Grippo, vice president of cloud platform at Workday, joined Scott Rosecrans, vice president of applications sales and strategic business development at Amazon Web Services (AWS), discussing the partnership between AWS and Workday.

The two companies have been allies for over a decade, building a partnership from a joint go-to-market perspective in the US, EMEA, and APJ. Now, AWS and Workday Extend have collaborated to create a native integration from Workday Extend to AWS, enabling developers to easily and securely leverage AWS platform capabilities in their Workday Extend apps. “AWS is excited about the growing global partnership with Workday,” shared Rosecrans. “The integrations we’ve developed with the Workday Extend team will help to both enhance the developer experience and accelerate innovation. We’re looking forward to continuing to co-innovate with Workday to develop new solutions that help our customers solve their business challenges.”

LESEN  6 Steps To Prioritize Natural & Paid Search In A Holistic Search Strategy

Developers competed in DevCon’s Hackathon, using the best of Workday and AWS’s AI and ML, storage and event-processing capabilities with storage and event-processing AWS AI services, AWS Lambda, Amazon S3, and Amazon EventBridge. Besides awarding prizes to the best and brightest, the Hackathon harnessed the power of developer collaboration and design thinking while underscoring the importance of “why.” 

When developers connect value to customers through what they do, they see the value of responsible artificial intelligence (AI) and their impact on the workforce more clearly. Development is not just coding and technical training. The work developers do make a difference. Tech solutions impact every part of an organization: HR, finance, marketing, sales and more. 

Think of it this way. If you engage people to build a ship, you don’t want to talk about the effort to build the ship. You want to show the builders that once the ship is built, everyone can cross the ocean and thrive on the other side. 

And on that other side is exactly where we find the best of workplace innovation where real-world problems get solved, efficiency is optimized, and user experience is enjoyable.

The Power of AI and ML

Workday’s co-CEO and chairman Aneel Bhusri kicked off the conference with Workday’s co-president, Sayan Chakraborty. They explored how AI and machine learning (ML) is changing the way work gets done. The power of AI and ML hits a high note on delivering strategic business results by allowing AI to handle more tedious tasks. In fact, Workday is opening up its platform ML capabilities in the form of APIs for developers that can be used in the applications they build to accelerate their business. Workday’s history in AI and ML technology is nothing new. The company has long taken a platform-first approach, accelerating the capabilities of what ML engineers and developers can build.

LESEN  Komplexe Daten mit Augmented Analytics vereinfachen

Updates on ML capabilities through the Workday platform include:

  • 13 new ML APIs through an early adopter program to accelerate business through developers’ work. 

  • The skills extract API, which helps developers extract relevant skills from text.

  • Document intelligence, which leverages Workday’s curated ML models to help users extract business relevant fields from business documents like supplier invoices, receipts, and generic documents.

  • The release of ML Forecaster for more intelligent forecasting.

There was a lot of buzz on how developers would create with these new apps, which illustrates an important point: DevCon is a rare gem among conferences. 

Tech leaders and business leaders need to talk. DevCon invites them to speak the same language, create with ingenuity, and understand each other’s worlds. Having tech and business leaders in one room is long overdue. As Workday’s fastest growing event—doubling in size since last year—developers who attended DevCon had the opportunity to explore new possibilities for innovation through in-depth sessions on capabilities like Workday ExtendWorkday Prism AnalyticsWorkday Adaptive Planning, Und Workday Journeys.

Bridging the gap enables developers to see the outcome of their efforts, not just the input of their efforts.

With HR and finance driving the work of many developers, large language models (LLMs) are one technique to solve problems. then demonstrated how the use of LLMs can be applied in Workday with Workday Prism Analytics. “With Prism, IT can bring in third-party data, combine it with Workday data, apply machine learning, and produce unique and powerful applications based on the output,” shared Chakraborty. 

The Power of Community

A community allows people to visualize how to do things faster and to learn from one other. This ecosystem inspires learning, networking, extending capabilities and imagination. 

Carl Eschenbach, co-CEO at Workday, talked about scaling businesses and helping customers navigate digital transformation in a conversation with Jay Wieczorkowski, Sr. director of cloud platform product management at Workday. Eschenbach, who comes with a deep understanding of Workday due to his board experience and close friendship with fellow co-CEO Bhusri, drew a straight line to culture and values, and emphasized the critical need for speed and agility across the company. 

LESEN  AWS and Zscaler expand relationship to help companies move to the cloud securely

Eschenbach discussed the need to innovate and move faster than ever before, with businesses moving from best of breed to best of platform. This creates opportunities for developers to tailor Workday for their needs.

He called developers “the drivers for the next chapter of growth for both Workday and their own businesses.” The developer community, he said, is key to readying companies for whatever changes come their way.

What Makes Workday’s Developer Community Different

One takeaway from DevCon is that workplace innovation is being redefined. With an aging workforce and rapidly changing technology and business models, modern organizations will rely more and more on internal skills to thrive.

Workday’s focus on developers is not a one-time event. Jim Stratton, Workday’s Chief Technology Officer, took the stage with Shane Luke, Vice President, AI and Machine Learning and Dave Sohigian, Global CTO. The three tech executives discussed their dedication to be expanding the range of features available to the developer community through the Workday Platform. Three areas Workday is committing to in the future are:

●  Embedded AI and ML

●  Platform interoperability

●  No/low-code development

By prioritizing these areas, Workday gives developers an opportunity to drive the innovative workplace. Luke said Workday approaches AI and ML technologies differently by focusing on data quality rather than just quantity and the use of federated learning. He shared Workday’s commitment to trustworthy AI and ML, ensuring transparency and ethical treatment. 

Tap into Workday’s free resources page with more than 2,000 videos, articles, demos, ebooks, webinars and more on a range of workplace innovation topics and solutions. And, if you haven’t already, put DevCon 2024 on your calendar. It could be the one event that amps up your power centers as the future of work unfolds.

Quellenlink

Behalten Sie im Auge, was wir tun
Seien Sie der Erste, der neueste Updates und exklusive Inhalte direkt in Ihren E-Mail-Posteingang erhält.
Wir versprechen, Sie nicht zuzuspammen. Sie können sich jederzeit abmelden.
Ungültige E-Mail-Adresse
Weiterlesen

TECHNOLOGIE

How Optical And Intelligent Character Recognition Works And Its Benefits

Veröffentlicht

An

How Optical And Intelligent Character Recognition Works And Its Benefits

Companies are using manual processes for Know-Your-Customer (KYC) verification, contracts, patient reports, order forms, and insurance claims.

For them, the paperless work environment is still a dream whilst digitization is getting on. 

Document processing technology has been around for a long time ago, it serves a single purpose or sticks to a certain document like invoice processing. Paper documents received by banks, healthcare providers, and insurance companies are stuck anywhere in the document bundle. Not only this, documents in different templates from different vendors, with non-standard field signatures, photographs, legal-sized or letter-sized, and rent orientation also stuck.  

What is Optical Character Recognition (OCR)? 

5 Open Source Frameworks to Build Your IoT Solutions

OCR is the procedure that converts image-to-text into a machine-readable format. For example, if an employee scans a receipt or a form, the software saves the process as an image file. Employees can’t use any text editor to edit, explore, or word count in the image file.

OCR struggles with distortions and non-standard document formats that lead to inaccuracies that hit productivity poorly and hinder straight-through processing rates. It captures just analog and singular moments that can’t be copied, shared, saved, or archived in the cloud. The information can’t be reassembled or broken down into pieces to build new stories and metadata is completely lost at this moment, as it were. 

LESEN  6 Steps To Prioritize Natural & Paid Search In A Holistic Search Strategy

Companies are looking for ways to capture metadata and have the potential to reach a paperless world through intelligent automation with OCR. That enables organizations to use end-to-end processing and capture through recognition to validation, classification, verification, and integration with EHR, ERP, SCM, CMS, and core platforms. 

How to Get Started with OCR?  

OCR is a very useful way to process diverse documents like PDFs, photocopies, images, handwritten documents, grainy photographs, and faxes with end-to-end encryption. The OCR technology identifies field extraction of relevant data like customer name and country. Moreover, systems also recognize them through different categories like rule, content, and image-based. 

Verification can also be done manually and IDP implementation changes the whole scenario of departments. There are strategies that organizations can utilize to ensure the successful implementation of advanced techniques. 

What is Intelligent Character Recognition (ICR)?

ICR is the most advanced way of optical character recognition. It is used to extract and interpret all types of information from scanned forms and documents like IDs, invoices, and contracts. It uses computer algorithms to process and recognize letters or numbers and scanned images. 

By using machine learning and natural language processing, it has the potential to cope with unstructured documents and those areas, where the information is not always in the same order or layout. Whereas OCR turns a scanned invoice into an undifferentiated character stream, ICR has the potential to identify which characters represent the company name, date, and amount.

LESEN  Google Says They Have Algorithms To Detect & Demote AI Plagiarized Content

How to Get Started with ICR?

Getting started with ICR requires hands-on training for CFOs and business analysts to understand this.  Employees that take part in document verification require training on how they can assist. Although it will open them for more skillful tasks or duties. Due to these challenges, a clear and concise structure is needed to implement ICR successfully. It means structured and unstructured change management and IT  or business functions alignment are crucial. 

ICR initiatives will be needed to align with automation which requires collaboration between the automation center of excellence and the central data capture team. It will be helpful to embed IR in the automation as this can streamline cross-linking. 

To monitor and measure the ICR solution’s effectiveness, KPIs will be needed to proactively define, identify and revise through the progress of implementation. Some of the best practices are providing robust skillful training with historical documents. Not only this validating the correctness of data labels that are used to train the models with documents available to train the model. 

Major Difference Between OCR and ICR

ICR is a part of OCR but the major difference is usually not set up to identify handwriting. It usually scans the paper and typed documents that turn into texts so they can be categorized and searched easily. 

LESEN  Meta Outlines Advances in Language Translation, a Crucial Consideration in Furthering Metaverse Connection

OCR-scanned texts can easily be copied and pasted while on the other hand, ICR technology focuses on printed or handwritten materials that use more complex fonts. 

How OCR and ICR Help Large Organisations? 

3 Technologies Every Digital Marketer Should Use

ICR and OCR help large organizations in diverse ways: 

Improve Data Accuracy and Fast Entry Speed

Manual data entry is labor-intensive, prone to human error, and costly. ICR and OCR support companies by digitizing data sources and eliminating manual data entry bases. 

Workflow Automation 

Businesses rely on high paperwork volume and scanning helps them to increase productivity and saves time. For Example, invoices and bills entered into workflow automation provide an instant pipeline from receipt to digital delivery to the required departments. 

Summing up

Physical documents can be damaged or lost, and an OCR and ICR offer a digital backup that takes a required space. The digital information can search quickly versus manually searching the documents in the storage area.

Quellenlink

Behalten Sie im Auge, was wir tun
Seien Sie der Erste, der neueste Updates und exklusive Inhalte direkt in Ihren E-Mail-Posteingang erhält.
Wir versprechen, Sie nicht zuzuspammen. Sie können sich jederzeit abmelden.
Ungültige E-Mail-Adresse
Weiterlesen

TECHNOLOGIE

90% of IT buyers have bought tech their infrastructure can’t support

Veröffentlicht

An

Cloud-Computing-Nachrichten

Pure Storage (NYSE: PSTG), a specialist in data storage technology and services, in partnership with Wakefield Research, has released a new report identifying the obstacles organisations across industries face in today’s digital economy while on the path to modernisation, and the investment priorities they are focused on to support business growth.

The new report, “IT Leader Insights: The State of IT Modernization Priorities and Challenges Amid Economic Headwinds,” reveals the need to evolve from legacy infrastructure in order to reap the full benefits of emerging technologies investments.

The survey, fielded among 500 IT buyers at companies of 500+ employees across the US and Europe, found that:

More Than Half of IT Buyers are Prioritising AI/Machine Learning and Sustainability Technology Investments: IT buyers have cited that their top investments planned for the next five years are AI/machine learning (52%) and sustainable technology (51%). While less than half (46%) predict they’ll invest in infrastructure automation or orchestration.

  • Nine in 10 IT Buyers Admit to Buying Tech Their Infrastructure Could Not Support: An overwhelming 90% of IT buyers stated that the pressure of their digital transformation agenda led them to buy technology their infrastructure could not support—and 74% say this resulted in being unable to deploy new technology to leadership’s full expectations.
  •  IT Buyers Continue to Feel Pressured to Invest in New Technologies: More than 3 in 5 IT buyers (62%) feel pressured all the time or often to make decisions on purchasing technology based on current needs without fully exploring the consequences of these decisions in the longer term. Seventy-six percent expect increased scrutiny on purchasing decisions over the next five years.
  • Nearly All IT Buyers Plan to Modernise IT Infrastructure to Drive Meaningful Outcomes: Nearly all survey respondents (99%) plan to modernise their IT infrastructure to better support future technology investments over the next five years, of which 73% are planning significant upgrades. The top areas of IT infrastructure in need of updates to better support adopting new technology include network and security (58%), data management or storage (52%) and data centre facilities (50%).
LESEN  How bias in AI can damage marketing data and what you can do about it

Digital transformation has accelerated at lightning speed, bringing an unprecedented need for IT modernisation. In the face of economic headwinds, IT teams are still focused on operationalising and supporting new digital initiatives, but with the added pressure of tighter budgets and lower staffing levels.

Amidst current inflation and recessionary concerns, IT teams must strike a complex balance between supporting critical IT services for their organisation and setting up a foundation on which future initiatives can thrive. With thoughtful investments, organisations can keep vital systems running efficiently while accelerating their ability to keep pace with digital transformation.

Shawn Rosemarin, VP, R&D and Customer Engineering, Pure Storage, said: “Investments in emerging technology such as AI and Machine Learning are exciting ventures for organisations interested in gaining a competitive edge, but without a modern IT infrastructure foundation the long-term success of these investments is at critical risk.

“The findings in our report show how agile IT leadership will need to continue to challenge legacy infrastructure solutions in order to simplify their core structure, in turn driving significant efficiency and flexibility to fuel their future.”

Want to learn more about cybersecurity and the cloud from industry leaders? Check out Cyber Security & Cloud Expo taking place in Amsterdam, California, and London. Explore other upcoming enterprise technology events and webinars powered by TechForge Hier.

  • Duncan ist ein preisgekrönter Redakteur mit mehr als 20 Jahren Erfahrung im Journalismus. Nachdem er seine Karriere als Tech-Journalist als Redakteur von Arabian Computer News in Dubai begonnen hatte, hat er seitdem eine Reihe von Publikationen zu Technik und digitalem Marketing herausgegeben, darunter Computer Business Review, TechWeekEurope, Figaro Digital, Digit und Marketing Gazette.

Tags: budget, Infrastructure

Quellenlink

Behalten Sie im Auge, was wir tun
Seien Sie der Erste, der neueste Updates und exklusive Inhalte direkt in Ihren E-Mail-Posteingang erhält.
Wir versprechen, Sie nicht zuzuspammen. Sie können sich jederzeit abmelden.
Ungültige E-Mail-Adresse
Weiterlesen

Im Trend

de_DE_formalDeutsch (Sie)