Connect with us

Big Data

AI models could help companies overcome human bias

Published

on

Machine learning algorithms can reinforce human bias, but representatives from HireVue Inc. and Kantar Millward…

Brown recently argued that they may also be able to remove our biases from processes.

HireVue CTO Loren Larsen said the AI models developed by the on-demand video interview platform company not only make the search for strong candidates more efficient, they also make it fairer for those applying.

The technology enables companies to scale the search process, meaning they can “take more chances and just let someone take [an interview] slot,” Larsen said at the recent Emotion AI Summit in Boston. By adding machine learning, HireVue is hoping to take things a step further and reduce human bias in the hiring process.

Take the example of how a candidate’s looks affect the job search. A HireVue data scientist developed an AI model to determine how much attractiveness might factor into hiring decisions. The model was trained on a public database of images and then was used to score attractiveness on a scale from one to 10.

“It turns out that if you got a seven or higher, you’re twice as likely to get hired than if you were a three,” Larsen said. That figure might be palatable if attractiveness equated to job performance, but, HireVue’s study couldn’t find a correlation between the two.

To that end, HireVue has striven to build AI models that can predict a job applicant’s potential performance — without a human in the loop. The models look for “traditional competencies,” according to Larsen, such as a candidate’s emotional awareness; negotiation skills; ability to collaborate, work with a team and learn.

HireVue’s AI models not only consider what’s being said by job candidates, but how it’s being said. They’re trained to factor in facial expressions and emotion — technology that’s powered by Affectiva, a software company spun out of the MIT Media Lab as well as the conference host.

AI models in advertising

At Kantar Millward Brown, a market research company based out of London, Affectiva’s software is helping make the case for more inclusive commercials. The company specializes in “advertising development work.” It helps clients understand how their ads are likely to be received by viewers and then finds ways to make them better.

“Some of that is done in what this audience may think of as a relatively old-school way: We show people the ads and ask them questions,” said Graham Page, executive vice president and head of global research solutions, at the summit.

Some of the work is done in a decidedly modern way. The firm films participants in a focus group as they watch an advertisement, and then it analyzes facial expressions and other  physiological data using Affectiva’s software “to understand the emotional response to the ad as it plays and what the key moments are that really resonated with people,” Page said.

For example, an analysis of advertisements done for Unilever, one of Kantar Millward Brown’s biggest clients, found that the ads categorized as “more progressive,” or more diverse, were 25% more effective than advertisements categorized as “less progressive,” or more stereotypical. And ads categorized as the least progressive were twice as likely to achieve the lowest scores on effectiveness, according to Page.

He described this study and others that have shown similar findings as “instructive” in that they help build a case for other businesses that “things like progressive advertisements are not only ethically the right thing to do, they’re also good for business,” he said.

‘IT departments suck’

IT’s reputation is still dubious, at least according to the VC panelists at the conference. When the moderator asked what advice the VCs could provide startups on how to sell to corporations, Krishna Gupta from Romulus Capital didn’t mince words: “IT departments suck.” He described integration as a rate limiter for many companies.

Janet Bannister, partner at Real Ventures in Montreal, suggested startups fret less about selling against other startups and more about selling against incumbents. She said large companies might understand that a startup can solve a problem better than the technology they’re currently using, but see the startup’s future as uncertain. “Having a strong use case, other customers using the product and great investors that will speak on the company’s behalf” may help assuage a large company’s concerns, she said.

Say what?!?

“Humans are unique. We’re awesome. Let’s get beyond that point and look at the attributes that we need in an artificial intelligence system that would enable us to trust it with more and more functionality. I think it’s a continuum. Just like ethics is a continuum. Morality is a continuum. … And I think we need to invite our machines into that continuum, that struggle, that wrestle that we’re in.” — Babak Hodjat, founder and chief scientist, Sentient Technology

“It’s kind of a tough time to think about how we encourage people to trust AI. And that’s particularly true given that some of the biggest businesses that use AI, particularly in the social sharing space, are at the absolute center of a massive crisis of trust.” — Graham Page, executive vice president and head of global research solutions, Kantar Millward Brown

Coinsmart. Beste Bitcoin-Börse in Europa
Source: https://searchcio.techtarget.com/news/252449410/AI-models-could-help-companies-overcome-human-bias

Big Data

If you did not already know

Published

on

SGAN google


The Generative Adversarial Networks (GANs) have demonstrated impressive performance for data synthesis, and are now used in a wide range of computer vision tasks. In spite of this success, they gained a reputation for being difficult to train, what results in a time-consuming and human-involved development process to use them. We consider an alternative training process, named SGAN, in which several adversarial ‘local’ pairs of networks are trained independently so that a ‘global’ supervising pair of networks can be trained against them. The goal is to train the global pair with the corresponding ensemble opponent for improved performances in terms of mode coverage. This approach aims at increasing the chances that learning will not stop for the global pair, preventing both to be trapped in an unsatisfactory local minimum, or to face oscillations often observed in practice. To guarantee the latter, the global pair never affects the local ones. The rules of SGAN training are thus as follows: the global generator and discriminator are trained using the local discriminators and generators, respectively, whereas the local networks are trained with their fixed local opponent. Experimental results on both toy and real-world problems demonstrate that this approach outperforms standard training in terms of better mitigating mode collapse, stability while converging and that it surprisingly, increases the convergence speed as well. …

Expert-Augmented Machine Learning (EAML) google


Machine Learning is proving invaluable across disciplines. However, its successis often limited by the quality and quantity of available data, while its adoption by the level of trust that models afford users. Human vs. machine performance is commonly compared empirically to decide whether a certain task should be performed by a computer or an expert. In reality, the optimal learning strategy may involve combining the complementary strengths of man and machine. Here we present Expert-Augmented Machine Learning (EAML), an automated method that guides the extraction of expert knowledge and its integration into machine-learned models. We use a large dataset of intensive care patient data to predict mortality and show that we can extract expert knowledge using an online platform, help reveal hidden confounders, improve generalizability ona different population and learn using less data. EAML presents a novel framework for high performance and dependable machine learning in critical applications. …

Multiple Independent Subspace Clustering (MISC) google


Multiple clustering aims at discovering diverse ways of organizing data into clusters. Despite the progress made, it’s still a challenge for users to analyze and understand the distinctive structure of each output clustering. To ease this process, we consider diverse clusterings embedded in different subspaces, and analyze the embedding subspaces to shed light into the structure of each clustering. To this end, we provide a two-stage approach called MISC (Multiple Independent Subspace Clusterings). In the first stage, MISC uses independent subspace analysis to seek multiple and statistical independent (i.e. non-redundant) subspaces, and determines the number of subspaces via the minimum description length principle. In the second stage, to account for the intrinsic geometric structure of samples embedded in each subspace, MISC performs graph regularized semi-nonnegative matrix factorization to explore clusters. It additionally integrates the kernel trick into matrix factorization to handle non-linearly separable clusters. Experimental results on synthetic datasets show that MISC can find different interesting clusterings from the sought independent subspaces, and it also outperforms other related and competitive approaches on real-world datasets. …

Fractal AI google


Fractal AI is a theory for general artificial intelligence. It allows to derive new mathematical tools that constitute the foundations for a new kind of stochastic calculus, by modelling information using cellular automaton-like structures instead of smooth functions. In the repository included we are presenting a new Agent, derived from the first principles of the theory, which is capable of solving Atari games several orders of magnitude more efficiently than other similar techniques, like Monte Carlo Tree Search. The code provided shows how it is now possible to beat some of the current state of the art benchmarks on Atari games, without previous learning and using less than 1000 samples to calculate each one of the actions when standard MCTS uses 3 Million samples. Among other things, Fractal AI makes it possible to generate a huge database of top performing examples with very little amount of computation required, transforming Reinforcement Learning into a supervised problem. The algorithm presented is capable of solving the exploration vs exploitation dilemma on both the discrete and continuous cases, while maintaining control over any aspect of the behavior of the Agent. From a general approach, new techniques presented here have direct applications to other areas such as: Non-equilibrium thermodynamics, chemistry, quantum physics, economics, information theory, and non-linear control theory. …

Coinsmart. Beste Bitcoin-Börse in Europa
Source: https://analytixon.com/2021/06/15/if-you-did-not-already-know-1423/

Continue Reading

Big Data

KDnuggets Top Blogs Rewards for May 2021

Published

on

KDnuggets Top Blogs Rewards for May 2021

We announce the winners of the first KDnuggets Top Blog Rewards Program.


Top Blog RewardsCongratulations to winners of KDnuggets Top Blogs Reward Program for May 2021, which will share the total amount of $2,000 (USD) according to the unique page views in the first 21 days starting with the day of publication.

  1. Vaex: Pandas but 1000x faster, by Ahmad Anis (original)
  2. Data Preparation in SQL, with Cheat Sheet!, by Stan Pugsley (original)
  3. A Guide On How To Become A Data Scientist (Step By Step Approach), by Aditya Agarwal
  4. Top Programming Languages and Their Uses, by Claire D. Costa (original)
  5. Essential Linear Algebra for Data Science and Machine Learning, by Benjamin Obi Tayo (original)
  6. Charticulator: Microsoft Research open-sourced a game-changing Data Visualization platform, by Josh Taylor

We started this program to encourage more high-quality and especially original (unpublished) contributions to KDnuggets.

For each month, starting in May 2021, we will determine 6 most viewed blogs published on KDnuggets that month (excluding blogs by KDnuggets Editors Matthew Mayo and Gregory Piatetsky), and will distribute $2,000 (USD) among the authors of these top blogs. If a blog has multiple authors, we will send the payment to the first author and it will be their responsibility to distribute it. If the author declines the monetary reward, the amount will roll over to the next month rewards.

While we appreciate and publish the reposts, the goal of this program is to get more great original publications, which will be rewarded at the rate of 3X the reposts. Note: An original submission first published on KDnuggets may be reposted elsewhere two weeks after KDnuggets publication, provided there is a link to KDnuggets blog and text “Originally published on KDnuggets”.

Here are the details on KDnuggets Top Blogs Reward Program .

To submit a blog to KDnuggets, please follow the submission guidelines.

Coinsmart. Beste Bitcoin-Börse in Europa
Source: https://www.kdnuggets.com/2021/06/top-blogs-rewards-may.html

Continue Reading

Big Data

Americans are curious about electric trucks, but low cost, durability come first – Reuters/Ipsos poll

Published

on

By Tina Bellon

(Reuters) – While most Americans expect electric vehicles eventually to dominate the nation’s roadways, many have greeted the first battery-powered models with skepticism, expressing concerns about the potential costs and inconveniences of owning such vehicles, a Reuters/Ipsos poll showed.

Automakers have already recognized the challenge and said they plan to market future battery-powered models by touting their performance and long-term cost savings over environmental benefits.

Vehicle manufacturers need to win over more rural Americans and Republicans, with the survey showing those groups express less enthusiasm than others for electric vehicles.

Vehicle electrification, particularly of the most popular large, fuel-intensive light trucks and SUVs, has become a central part of U.S. efforts to reduce greenhouse gas emissions to combat climate change. In the Reuters/Ipsos national poll of around 1,005 adults from June 3-4, 67% of all respondents agreed in a bipartisan consensus that pickup trucks are the American way of life, underscoring the focus on electrifying those vehicles to reach climate goals.

Rural consumers in so-called red states or Republican strongholds are a key demographic for pickup trucks, the auto industry’s most profitable vehicle segment, challenging automakers to navigate an electric future without alienating those customers.

Several automakers, including Ford Motor Co and General Motors Co, are planning to release all-electric trucks in the coming years. But no electric pickup truck is yet available, which means consumer attitudes about them are largely based on second-hand knowledge.

Graphic-What do you want in a truck? https://graphics.reuters.com/AUTOS-ELECTRIC/SURVEY/qzjpqrgynvx/chart_eikon.jpg

A majority of 65% agreed that electric vehicles are the future of the auto industry, the poll showed. While there were clear divisions along party lines – with 53% agreeing among Republicans, compared with 78% among Democrats – the issue is less partisan than other contentious U.S. political topics.

Nearly half of Americans – 46% – believe EVs are not worth the cost, the survey showed, with opinions again reflecting a political divide. EVs have a higher sticker price than comparable gasoline models, largely due to costly batteries.

When asked to consider buying a new truck, only a minority said they would search for one that minimizes its impact on the environment. Thirty-eight percent said they would look first for an efficient truck that costs less to operate. Another 34% said they wanted a durable truck, and only about 19% said they wanted an environmentally friendly one.

PERFORMANCE BENEFITS

Jim Richman, a window installer in Austin, Texas, who owns a 2018 Ford F-150, said he is also concerned about the reliability of electric trucks and the hassle of charging them.

“I just want to know my truck is ready whenever I need it, and also for longer trips,” Richman said, echoing a sentiment among a majority of survey respondents who said EVs generally were inconvenient because of their need for charging.

The auto industry is trying to alleviate those concerns with technological innovations, including reductions in battery costs and longer driving ranges.

GM, which plans to start selling an all-electric version of its Hummer pickup by the end of next year, said more consumers will consider EVs as they become familiar with battery-powered vehicles. GM is also trying to emphasize economic benefits of going electric.

“We have a tremendous opportunity to help consumers understand that benefits of driving electric aren’t just related to the environmental benefits, but performance and capability that EVs offer,” said Darryll Harrison, a GM spokesman.

Researchers expect EVs to reach price parity with fuel-powered counterparts around 2030, and EVs also provide long-term cost savings due to lower maintenance costs and the lower cost per mile of refueling with electricity instead of gasoline.

Ford said its research had shown consumer demand for an e-truck, especially among younger truck owners. The company plans to release its all-electric F-150 Lightning truck next year, for which it has received some 100,000 reservations to date.

“By 2030 we expect 40% of our global mix to be fully electric and we’ll achieve that by electrifying our most popular name plates,” said Darren Palmer, Ford’s North America EV general manager.

Chris Sherman, general manager of a Ford dealership in rural Paris, Texas, said he has been surprised by the positive response to the Lightning so far.

“These are pretty much diehard gas and diesel pickup owners,” Sherman said. “Lot of people here are blue-collar, and they tow utility trailers. They want to know how much of a day’s work they can do with their truck.”

Sherman said he had so far received 10 reservations for the electric truck and that he fields up to 40 phone calls per day about it.

The Reuters/Ipsos poll was conducted online, in English, throughout the United States. It gathered responses from 1,005 adults, and the results have a credibility interval, a measure of precision, of about 4 percentage points.

(Reporting by Tina Bellon in Austin, Texas; Additional reporting by Paul Lienert in Detroit; Editing by Joe White, Chris Kahn and Matthew Lewis)

Image Credit: Reuters

Coinsmart. Beste Bitcoin-Börse in Europa
Source: https://datafloq.com/read/americans-curious-electric-trucks-low-cost-durability-come-first-reutersipsos-poll/15475

Continue Reading

Big Data

GM’s self-driving car unit Cruise to access $5 billion credit line

Published

on

BERKELEY, California (Reuters) -General Motors Co’s majority owned Cruise self-driving car subsidiary said on Tuesday it would gain access to a multi-year $5 billion line of credit with GM’s finance arm to fund its purchase of Origin vehicles starting in early 2023.

“This bumps up Cruise’s total war chest to over $10 billion as we enter commercialization,” Dan Ammann, chief executive of the Cruise unit, which also counts Softbank and Honda as minority investors, said in a blog posting.

Self-driving startups are racing to build warchests to develop and scale up technology, an expensive and time-consuming process.

This past month, General Motors began assembling the first of about 100 pre-production Cruise Origin all-electric vehicles to be built this summer for validation testing, Cruise said.

Cruise, which has been testing driverless GM Chevy Bolt electric vehicles in San Francisco, plans to commercialize the technology in the densely populated city “in months, not years,” spokesman Ray Wert said.

Reuters last month reported that Cruise and rival Waymo, a unit of Alphabet Inc’s, have applied for permits needed to start charging for autonomous rides and delivery in San Francisco.

(Reporting by Hyunjoo Jin; Editing by Edmund Blair and David Gregorio)

Image Credit: Reuters

Coinsmart. Beste Bitcoin-Börse in Europa
Source: https://datafloq.com/read/gms-self-driving-car-unit-cruise-access-5-billion-credit-line/15474

Continue Reading
Aviation2 days ago

Delta Air Lines Flight Diverts To Oklahoma Over Unruly Off-Duty Flight Attendant

Crowdfunding5 days ago

April/May 2021 Top Campaigns

Esports3 days ago

Lost Ark Founders Pack: Everything You Need to Know

Blockchain5 days ago

Crypto Fund Manager Says Bitcoin ETFs to be Approved By 2022

Fintech5 days ago

PayPal launches PayPal Rewards Card in Australia

Aviation2 days ago

Spirit Airlines Just Made The Best Argument For Lifting LaGuardia’s Perimeter Rule

Cyber Security4 days ago

Data Breach that Impacted Both Audi of America and Volkswagen of America

Cleantech4 days ago

Tesla Model S 420 Plaid Is The Best Car In The World (But Not For Me)

Energy4 days ago

Industrial robots market in the automotive industry | $ 3.97 billion growth expected during 2021-2025 | 17000+ Technavio Research Reports

Energy4 days ago

Daiki Axis Co., Ltd. (4245, First Section, Tokyo Stock Exchange) Overview of Operating Performance for the First Three Months Ended March 31, 2021

Fintech5 days ago

Stripe launches Stripe Tax to simplify global tax compliance for Australian businesses

Blockchain4 days ago

Blockchain technology can help to protect sensitive information

AI5 days ago

Ransomware Incidents Surging; Cybersecurity Experts Scramble to Respond   

Blockchain4 days ago

Blockchain technology can help to protect sensitive information

Esports5 days ago

V-Union cards officially revealed for the Pokémon TCG, includes Mewtwo, Zacian, and Greninja

Blockchain4 days ago

DCR Technical Analysis: Look for Support Levels of $130.13 and $126.01

Blockchain4 days ago

State-Chartered Banks in Texas to Provide Crypto Custody

CNBC5 days ago

From green energy to cybersecurity, Citi names ‘unstoppable trends’ that investors can jump on

Esports5 days ago

New product details and art shared for Pokémon TCG set Evolving Skies

Esports5 days ago

LEC Power Rankings: 2021 Summer Split Preseason

Trending