One of the biggest misconceptions about 3D printing – the notion that it doesn’t suit the production of goods in high volume – has been put to bed by brands such as Adidas, which has turned heads in the footwear industry with its striking midsole designs. Virtual reality (VR) is similarly misunderstood and sometimes dismissed as a gimmick. But put VR and 3D printing together, and you have a rapid prototyping dream team that’ll be the envy of your competitors.

And don’t think that you need to spend north of US $3000 on an Apple Vision Pro to reap the rewards. The benefits of combining VR and 3D printing can be realized with an affordable headset such as the Quest 3 – released in 2023 – or even a Quest 2, which is now available at a lower price point and with a software speed up.

Iterate faster

A big fan of having designers use VR to iterate designs in 3D is Jared Goldman – VP of Design at New Balance (another footwear company at the bleeding edge of manufacturing). And there are some compelling reasons to back up his thinking.

READ NEXT

3D printed shoes – footwear signals future manufacturing trend

By Goldman’s estimation, physical samples have a turnaround time of around 45 days – once you’ve added up sending the tech package, production of the prototypes in Asia, and shipping of the shoes back to the US. However, it’s possible to produce photoreal samples of the same designs in seven days, or maybe less, in a virtual environment.

Decision makers can see the shoes from all angles in a virtual environment and have confidence in their feedback that compares with handling physical samples. Plus, headset-wearing team members can collaborate easily and in real-time, no matter where they are across the globe – provided that they have an internet connection.

“The better you can express your idea, so that somebody who is a non-designer can understand it, the more successful you’re going to be,” Goldman comments in a case study shared by Gravity Sketch – a developer of VR design software that is increasingly becoming part of 3D printing product workflows.

Example of a VR and 3D printing toolchain

• Gravity Sketch – for virtual product creation
• Blender – to add materials appearance for final design validation
• Ultimaker Cura – for slicing and 3D printing production preview
• 3D printers – to produce the finished goods

Today, there are numerous online tutorials showing how to combine VR and 3D printing to go from first idea to finished product. Beginning with a rough sketch created using handheld VR controllers, the next step is to add virtual form to the digital design, with lighting effects making the output appear all the more realistic.

Slicing software will prepare the model for 3D printing, where it’s possible to adjust support material and preview the build to double-check that all’s well before committing fabrication time and materials to the job.

And it’s not just about keeping sneakerheads happy with a stream of new and exclusive trainers, VR and 3D printing can be combined to deliver progress in areas such as medical visualization too.

Materialise – a pioneer in the field of 3D printing – points out that the number of hospital-based 3D printing facilities has increased exponentially to help with tasks such as patient-specific surgical planning.

READ NEXT

Fungi, coffee, and pineapples – rethinking device materials

On TechHQ, we’ve written about how beneficial VR can be to medical training and upskilling hospital staff. And 3D-printed anatomical models take that tactile experience to the next level.

Construction is another area that’s taking a keen interest in additive manufacturing – this time using giant 3D printers to build homes layer by layer. Here, VR can be used by designers and home buyers to experience what it’s like to move around a digital render of the building prior before nozzles start squirting out cement.

“We see this tool as an exciting way the get clients, designers, and contractors up to speed on the inner workings of how a 3D construction printing project actually works,” commented Marchant Van Den Heever, CTO of HTL.tech – a distributor of 3D construction printing technology in the UK and Ireland.

The post Rapid prototyping dream team – VR and 3D printing appeared first on TechHQ.

Visualizing the future as one where humans do less manual and repetitive work and robots do more, depends on finding an efficient way of teaching machines to perform such tasks. Ideally, the skills transfer process would generate rich data and be fast and cheap to carry out, but coming up with a method that ticks all of those boxes has proven to be difficult – until now. Hitting that sweet spot appears to be a pair of GoPro-equipped robot gloves developed by researchers in the US, which – according to video footage – could provide an easy way of training robots to do all kinds of things.

What’s more, all of the universal manipulation interface know-how has been open-sourced, including the 3D printing instructions for making the handheld robot gloves. As photos reveal, the soft finger design is capable of gripping a raw egg securely without breaking the shell.

To begin the skills transfer process between human and machine, users put on a pair of robot gloves and carry out the target task multiple times to build a training dataset. Don’t be discouraged by the need for repetition, as the results can be generalized to similar scenarios – using a so-called diffusion policy that has been shown to outperform existing state-of-the-art robot learning methods – which saves time later on.

READ NEXT

Cube warehouse robots square up against humanoid designs

Adding to the appeal, those same results can be used by different models of robot – provided that the unit can be fitted with duplicates of the robot gloves. In the demonstrations given by the team, whose members are based at Stanford University, Columbia University, and Toyota Research Institute, robots are taught how to place an espresso cup on a saucer and even wash up dirty plates.

Key to the success of the approach is the use of GoPro cameras – one on each of the robot training gloves and one on each of the grippers in the robot-mounted setup. The cameras feature fish eye lenses to capture a wide field of view, gathering large amounts of detail from the scene, and include inertial measurement units (IMUs) – to enable pose tracking.

The team makes sure that all of the data feeds are latency-matched, which means that robots can carry out two-handed tasks correctly and perform actions such as throwing objects with high precision. Also, there’s a one-off mapping step that uses a visual code to help with simultaneous localization and mapping (SLAM).

If sufficient numbers of people join in, robots could quickly be taught to do many common industrial tasks using the open-sourced robot gloves – and that knowledge shared. Currently, robots are often taught through teleoperation, which can be a slow process. The wearable teaching grippers, on the other hand, provide a much speedier option and are more instinctive to use.

READ NEXT

Helping robots learn: GPT-3 tool descriptions add value

“By recording all information in a single, standardized MP4 file, UMI’s data can be easily shared over the Internet, allowing geographically distributed data collection from a large pool of nonexpert demonstrators,” writes the group in its paper – ‘Universal Manipulation Interface: In-The-Wild Robot Teaching Without In-The-Wild Robots’ – which is free-to-read on arXiv.

Timing the robot training process, the researchers found that using their universal manipulation interface was around three times faster to use than teleoperation. Also, the learning framework was shown to be tolerant to big changes in lighting conditions, and other interference.

For example, robots trained using the gloves can continue performing their tasks even if their base is moved or humans perturb the scene in other ways – such as adding more sauce to the dirty plates.

The dishwashing task is noteworthy as it’s what’s termed an ultra-long horizon task from an automation perspective, with the success of each step dependent on the previous one. Here, the robot needs to perform seven actions in sequence – turn on the faucet, grasp the plate, pick up the sponge, wash and wipe the plate until the ketchups are removed, place the plate, place the sponge, and turn off the faucet.

Can we collect robot data without any robots?

Introducing Universal Manipulation Interface (UMI)

An open-source $400 system from @Stanford designed to democratize robot data collection

0 teleop -> autonomously wash dishes (precise), toss (dynamic), and fold clothes (bimanual) pic.twitter.com/BBG4S97B1Q

— Cheng Chi (@chichengcc) February 16, 2024

Given the apparent success of the approach, regular dishwashing appliances may face some competition from two-armed robots in the future – and it get’s you thinking about other jobs that robots could do around the home.

The post GoPro-equipped robot gloves teach robots new tricks appeared first on TechHQ.

With season 6 of Netflix’s hugely successful Drive to Survive hitting screens on Friday, motorsport fans will be tuning in to reminisce about the highs and lows of the 2023 F1 World Championship. But what viewers may not realize is just how much can be learned more broadly from watching racers driving around a track. Motorsport, it turns out, has lessons for businesses of various types in understanding how employees will interact as company fortunes rise and fall.

“Sports – particularly motorsports – can be a good proxy for several other industries as they are extremely competitive: if you don’t perform and progress you may be out,” comments Hans Frankort – a strategy expert at Bayes Business School – City, University of London. “Workers in sectors such as consultancy and financial services face similar pressures.”

[10/12] This is the first study published in #management that uses @MotoGP data and analyzes hundreds-of-thousand overtakes between 2004 and 2020 as riders’ competitive actions. We have also used qualitative insights from top teams such as @ducaticorse . pic.twitter.com/dEtoBPy9cK

— Paolo Aversa (@aversapaolo) January 18, 2024

Frankort, together with the other authors of the study – dubbed ‘Revving up or backing down? Cross-level effects of firm-level tournaments on employees’ competitive actions’, used overtake data from riders competing in MotoGP from 2004 to 2020 to generate motorsport lessons for business leaders.

“Our findings reveal how riders adjust their internal and external overtakes based on their team’s competitive threats and opportunities, and on the relative resource endowments of the teams supplying such threats or opportunities,” writes the team in its paper.

READ NEXT

Sports tech: iPhones play to win

Translating this behaviour from the racetrack to the office, the experts note how ambitious workers will change their approach depending on whether their employer is doing well – in other words, leading the competition – or finding business conditions more challenging.

Considering the MotoGP data, teammates were less likely to overtake each other when the team as a whole was struggling. “If a firm is facing threats, such as losing market share to smaller rivals, workers may feel that infighting is poor form,” said Frankort. “Instead, they would focus on competing against rival firms.”

More motorsport lessons for business

There are other observations too – for example, replacement riders (whom the researchers liken to agency workers) are keen to challenge their teammates when the team is doing well, and all competitors when their employer is struggling. The explanation given is that those without a permanent contract will try much harder to impress than riders and drivers embedded within a team.

So how do all of these insights help CEOs and other senior management become better leaders? One of the strong takeaways is the link between how a company’s performance is pitched to employees and the competitive actions of that firm’s staff.

If employees believe that things are going well they may be more tempted to poach a colleague’s clients and position themselves as being the engine of success, according to the study’s authors.

READ NEXT

Sports industry explores winning ways with tech

Motorsport series such as F1 and MotoGP are often described as pathfinders for new technologies that will one day end up in road cars or motorcycles, but they have also provided numerous case studies for business leaders. For example, Paolo Aversa – one of the authors of the MotoGP study – has made a career of using sports data to advance management theory.

Over the years, Aversa has chaired multiple ‘Competing in turbulent environments: Lessons from Formula One’ events, which are soon sold out. The seminars, some of which are still available to watch on YouTube, bring together race car designers, motorsport CEOs, and lecturers in management and business strategy.

Tech firms such as Netflix have also done well from F1 and one of the motorsport lessons learned is that drama on and off track makes for great entertainment. On paper, the idea of having a thousand people produce two cars that drive in circles sounds like it’d be a hard sell to viewers, but the popularity of Drive to Survive tells a different story.

And motorsport fans will need no reminding that F1 pre-season testing gets underway today in Bahrain, where teams and fans will get a first impression of the pecking order for the 2024 season.

The post Motorsport lessons: what can tech CEOs learn from MotoGP & F1? appeared first on TechHQ.

• A recent report shows the rapid changes IT infrastructure is undergoing.
• Legacy tech can slow down an organization’s whole IT modernization program.
• IT leaders reveal the infrastructure problems they’re having – and what they forecast for the future.

Legacy infrastructure is a sustainability issue for IT professionals, who rank energy efficiency and sustainability as important, but don’t feel confident about meeting their green targets.

All this and more has been revealed by research by Daisy Corporate Services on legacy tech with a focus on the cloud. Researchers heard from more than 250 senior IT decision makers at large private and public sector organizations in the UK, uncovering what they consider to be the key IT infrastructure challenges facing them and their teams – and how to overcome them.

The report provides insights into IT decision makers’ challenges including the global IT talent shortage, the drivers for IT outsourcing and factors for on-premises vs. hosted IT infrastructure.

75% fully (24%) or partly (51%) outsource their organization’s IT infrastructure management to a managed service provider. Partly due to the global talent shortage, IT decision makers felt that their organizations lack the necessary skills and expertise to stay in-house.

If that wasn’t enough, the report further details predictions, barriers and benefits of moving more of the IT estate to the cloud and looks at IT budgets, investments, and the impact of and opinions on new tech like AI – thought it’d been left out?

Over the last five years, businesses have had to manage the significant evolution of business systems and IT landscapes, with the pandemic a driver of that change. Increasing competition and globalization have affected the operating environment, which has further changed as a result of political and supply chain instability.

The increasing range of technologies and applications was in joint first place for contributing most to IT complexity alongside cloud services (57% each). Full visibility into infrastructure performance is made difficult by both of these factors.

90% of IT decision makers say building, managing and maintaining their IT landscape has become more complex and simplifying IT infrastructure is a priority for 89%.

Moving your IT infrastructure to the cloud – an unstoppable trend?

Despite that, though, the benefits of cloud use are obvious, so it’s increasing.

The ability to harness the cloud is dependent on the existing IT landscape at a company, with the biggest barrier to success being that the greater complexity of existing IT infrastructure, the greater is the work needed to migrate to the cloud.

Another significant threat to organizations looking to move more of their IT estate to the cloud are data security concerns.

Too often, organizations find they’re using a cloud hybrid system by accident, not design. Which means falling at the first hurdle in a world where it’s vital that an organization’s infrastructure is secure and able to dynamically adjust to workload demands.

The multitude of different environments and tools in use make it hard for organizations already struggling to gain the end-to-end visibility necessary for consistent security and reliability.

There is a strategic impact from sustainability as the issue becomes a business differentiator; the importance of green policy use means that consumers look to a company’s sustainability as a major factor of their appeal.

The focus on corporate and social responsibility filters through to IT teams as businesses turn their attention to the sustainability and efficiency of their IT operations.

The biggest hurdle to a clean, green transition is legacy and on-premise infrastructure: legacy tech was described by 63% of IT decision makers as a “sustainability nightmare.” 86% say sustainability and efficiency is important to IT operations, and 84% that their organization has IT sustainability targets in place – but only 51% of them are “very confident” they’ll meet them.

And it isn’t just contributing to a difficult transition to greener systems: legacy hardware contributes to 37% of organizations’ overall IT power consumption.

Spending is also coming under pressure as the global economic slowdown means businesses are having to adapt for survival – driving efficiencies is one tactic for survival that has huge impact on IT teams.

IT leaders are under pressure to reduce expenditure, many resorting to cloud services to turn capital infrastructure expenses into opex costs.

READ NEXT

Protecting critical national infrastructure through secure cloud

“Sustainability is a vital component of any modern business, and IT departments have a growing role in helping the wider organisation achieve green targets. But legacy technology is a cause for concern among IT teams, with ageing equipment still contributing significantly to power consumption,” comments Andy Bevan, head of propositions and strategy consulting at Daisy.

“Organizations can benefit from the sustainability features of their cloud providers but are being held back by the challenges of migrating their legacy hardware. Here is where modern hybrid cloud platforms can help bridge the gap between on-site infrastructure and cloud to deliver performance and sustainability benefits.”

Many IT leaders are being asked to re-evaluate their IT spending. More than two thirds (69%) of survey respondents describe the pressure to reduce IT capital expenditure as “significant.” Many are looking to consumption-based pricing to reduce ongoing costs and increase flexibility by paying only for what they use – scaling up and using more resources at peak times, and then reducing resources and costs again afterwards.

Legacy infrastructure is a headache in this process though, as its maintenance and support incurs significant cost. New technology is the solution for some IT managers, with many expecting to see the benefits of using artificial intelligence and automation in areas like service and performance management.

Bevan adds, “At a time when IT leaders are under pressure to reduce capital expenditure many organizations are still incurring significant maintenance and support costs on their legacy hardware. By moving to the cloud and a consumption-based pricing model, organizations can reduce ongoing costs and increase flexibility by paying for what they use. For cost-constrained IT departments this should be their nirvana.”

The full research from Daisy, Faster, greener, cheaper – dealing with IT infrastructure complexity in a hybrid cloud world, is available here.

The post Legacy IT infrastructure is a sustainability nightmare appeared first on TechHQ.

• There’s a significant disparity between the avilability of tech jobs in California and the number of young people completing computer science education.
• Bills are in progres to make computer science education mandatory to graduation.
• BUt how much will htat actually shift the dial or solve the problem?

Home of Silicon Valley and a technology school with just 4% acceptance rates, California is a hive for computer science. Nearly 1.5 million people in the state work tech jobs and its tech workforce is the largest among all the states by a wide margin.

While California’s tech industry delivers an economic impact of $536 billion and there are an estimated 55,868 tech business establishments in the state, education for the younger generation doesn’t reflect the reputation to which the state is entitled.

Just five percent of the 1,930,000 high school pupils in California are enrolled in a computer science course. In fact, despite being a step ahead in technology employment rates, California lags behind the national average, and behind 40 other states, in the percentage of high schools offering  computer science education.

One Slashdot user noted that the imagery used in the announcement was “verbatim” lifted from Code.org.

To remedy this, at the beginning of 2023 legislation was introduced by Assemblymember Marc Berman to guarantee access for all high school students in California. Far from being the Silicon Valley state’s step above the rest of the country, AB 1054 aligned California with 27 other states that already required public high schools to offer a computer science education.

Furthering efforts on February 6 this year, at a press conference Berman was joined by State Superintendent of Public Instruction, Tony Thurmond, to unveil a bill that, if passed, would require every public high school to teach computer science.

By the 2030-31 school year, the legislation (AB 2097) proposes, computer science would be established as a requirement for high school graduation.

“It is critical to equip our students with the skills they need to enter the 21st century workforce and succeed in our digitally driven world,” said Berman.

“We owe it to our students to teach them the fundamental skills they need to succeed in tomorrow’s economy – and that starts with having access to a computer science education.”

As of January 2023, California has 45,245 open computing jobs that have an average salary of $153,544, yet there were only 9,339 graduates in computer science in 2020. Indeed, California has the highest number of open computing jobs in the country.

“Too many students who grow up in the shadows of global tech companies are not gaining the skills they need to one day work at those companies,” said Berman. “Not only will AB 2097 help provide the workforce needed for California to remain competitive with other states and other nations, but it is also crucial in closing the existing gender and diversity gaps.”

READ NEXT

Does AI have a place in schools?

“If we truly value equity in our schools, we need to ensure all students have access to computer science education.”

Although female students make up 49% of the school population in California, only 30% of computer science students are female. Does making it a required topic of study open the door to more girls?

Sure, if it’s alongside the right messaging – the fact that computer science doesn’t just set students up for careers in Silicon Valley or tech more broadly, but rather a life that will inevitably be more dominated by technology than it’s ever been before.

It’s true that only 34% of high schools that serve high proportions of Black, Indigenous, Latinx, and Pacific Islander pupils offer computer science courses, compared to 52% of schools serving a greater proportion of White and Asian pupils.

Again, though, simply enforcing that students can’t graduate without passing computer science won’t change much more than metrics. Yes, 100% of those schools will offer computer science education, but who will make sure that students have the resources outside of school to successfully engage in the lessons?

How much algebra do you remember from high school?

The announcement of the Bill comes less than two weeks after the CEO of nonprofit Code.org, Hadi Partovi, was a keynote speaker at the Association of California School Administrators Superintendents’ Symposium.

Co-founder and CEO Hadi Partovi at the Code.org 10th Anniversary Gala.

Code.org has backing from several tech giants and advocates for computer science education. Partovi’s goal is to make computer science a high school graduation requirement in all 50 states by 2030 – so California’s proposed bill is a good start.

Back in an October Facebook post, Berman noted he’d partnered with Code.org on legislation in the past and hinted that something big was in the works on the K-12 CS education front for California.

“I had the chance to attend Code.org’s 10th anniversary celebration and chat with their founder, Hadi Partovi, as well as CS advocate Aloe Blacc. They’ve done amazing work expanding access to computer science education… and I’ve been proud to partner with them on legislation to do that in CA. More to come!”

More did come, and no doubt this is just the beginning. What remains to be seen is whether enforcing computer science education will be enough to shift the public attitude – tech companies might be feeling some hostility in California at the moment.

Waymo certainly isn’t welcome in San Francisco!

happening NOW in SF. Waymo car vandalized & lit on fire @sfchronicle pic.twitter.com/OEZYFiy6mv

— Michael Vandi (@michael_vandi) February 11, 2024

Assemblymember Berman has strong history on computer science education in California.

The post Kids must code in California appeared first on TechHQ.

• Employee mental health appears unaffected by wellness apps.
• Unsurprisingly, better shcedules, more flexibility and more money work a treat.
• Wellbeing research appears to debunk the actual benefit of corporate wellness to staff.

READ NEXT

Technology that might actually help your mental health

The billion-dollar industry of employee mental health services has flourished as a way for businesses to signpost that they care about their staff.

Often billed as the topmost benefit on job listings, digital wellness solutions, mindfulness seminars, sleep apps and other packages that scream “your mental health matters!” is one of the first things a new employee will be presented with.

But are these anything more than a point of pride for the HR department? A British researcher who analyzed survey responses from 46,336 workers at companies offering such programs found that those who participated were no better off than their colleagues who skipped the Monday morning resilience workshop.

Not one of the interventions that companies selflessly offer – apps, coaching, relaxation classes, courses in time management – had any positive effect.

Employee mental health bettered by doing charity work

The study, published in the Industrial Relations Journal, explored the outcomes of 90 different interventions – individual-level approaches – and found that only one had marked positive impact on employees: given the opportunity, the wellbeing of workers who did charity or volunteer work seemed to improve.

“It’s a fairly controversial finding, that these very popular programs were not effective,” said William J. Fleming, the author of the study and a fellow at Oxford University’s Wellbeing Research Center.

READ NEXT

Stretching the salary: tech that gets your employees more from their money

“If employees do want access to mindfulness apps and sleep programs and wellbeing apps, there is not anything wrong with that,” he said. “But if you’re seriously trying to drive employee’s wellbeing, then it has to be about working practices.”

Is it a surprise that analysis suggests the key to bettering employees’ mental health is a focus on “core organizational practices” including schedules and pay?

Researchers say the findings come as no surprise but do call into question practices that have become common across almost all job sectors.

“Employers want to be seen as doing something, but they don’t want to look closely and change the way work is organized,” said Tony D. LaMontagne, a professor of work, health and well-being at Deakin University in Melbourne, Australia.

Who’d have thunk that the corporate wellness services industry didn’t have staff’s best interests at heart? Employee mental health has entered the corporate consciousness, certainly, but by now it represents a money saving strategy rather than a staff care plan.

That’s not to say that the HR department or company execs are uncaring, or don’t understand the impact of mental wellbeing on the workplace; work as it stands isn’t ever going to foster wellness because it ultimately doesn’t center on the human – it’s all about the money.

Investing in the interventions offered by wellbeing apps and seminar leaders is cheaper than increasing staff time off or, heaven forbid, wages.

Less pessimistically, the wellness package option is sold as a money-saving strategy that will improve worker health and productivity.

According to Adam Chekroud, a co-founder of Spring Health and an assistant professor of psychiatry at Yale, a blanket dismissal of workplace interventions risks “throwing the baby out with the bathwater.”

A 2022 study tracking 1,132 workers in the US who used Spring Health, a platform that connects employees with mental health services like therapy and medication management, found that 69.3% of participants’ depression was improved. They missed fewer days of work and reported higher productivity.

Babies in the bathwater?

“There is recent and highly credible data that things like mental health programs do improve all those metrics that [Dr. Fleming] mentions,” Dr. Chekroud said. “That’s the baby you shouldn’t be throwing out.”

Any attempt to take the employee mental health more seriously is a step in the right direction – even if only a teeny little one, taken by an Inner Child.

The post Mental wellness apps offered to staff don’t help appeared first on TechHQ.

Ever wondered what’s happening under the hood of today’s powerful chatbots? Having a better understanding of how large language models (LLMs) work not only shines a light on what’s possible, but also helps to pinpoint their limitations (and see through the hype). And a great place to start that journey is to learn from experts in applied machine intelligence, such as Mascha Kurpicz-Briki – who has just written a book on demystifying LLMs.

Her guide to how LLMs work steps through all of the need-to-know concepts underpinning natural language processing. And by the end of the journey, readers will have a clear understanding of what it takes for computers to process written text.

Machine learning enables AI

The path begins with machine learning – looking at how computers can learn to solve tasks involving similar, but previously unseen data – before moving into the world of deep learning. Here, readers are introduced to the fascinating fundamentals of neural networks and get to dive into the topic of word embeddings – representing human language as a series of mathematical vectors.

READ NEXT

Like the sound of private, air-gapped LLMs? Welcome to generative AI on CPUs.

A key takeaway here is that words of similar meaning have vectors that are closer together, which chimes with the linguistic notion that ‘you shall know a word by the company it keeps’ – attributed to J.R. Firth. And neural networks take this concept to the next level.

“Higher-dimensional vectors help to capture different properties of the words and thus improve the quality of the relations between the word embeddings,” writes Kurpicz-Briki.

Rather than learning the meaning of the word directly, algorithms generate those dictionaries of word vectors by recognizing common words that often appear together in sentences, the author points out. And this hints at the statistical marvel that is today’s chatbot.

By this point in the multi-stage tour into how LLMs work, readers can expect to be well-equipped with a high-level view of language models, as well as having an understanding of part-of-speech tagging and how word dependency is established.

One of the big advantages of LLMs, as Kurpicz-Briki highlights, is that their next-word predictive power considers not just how two words relate to each other, but also acknowledges ‘previous states’. And the fact that chatbots can consider other words in the sentence before predicting what’s likely to come next is a big tell on how LLMs work.

Reaching the heart of the book, the author considers how clever context mechanics have improved model output and tackles the topic of encoders and decoders.

READ NEXT

Beach reads for IT pros – summer 2022

Taking the sentence: ‘There is a field of strawberries, and it is so beautiful!’, as an example, she explains how applying self-attention boosts the model weight for field as it is effectively referred to twice.

“Using this mechanism of self-attention, we do not lose track of words referring to other words that have appeared previously in the sentence,” writes Kurpicz-Briki.

Transformer-based models enable contextualized word embeddings. For example, orange can appear alongside other fruits in the vector space. But the word can be mapped with colors too, recognizing that it has different meanings, which are encoded accordingly by considering whole sentences.

Encoders and decoders

Encoders can be trained using word masking, and – once complete – they pave the way for next-sentence prediction, which begins to paint a picture of how decoders work.

At a high-level, finding information that is similar, in terms of its vector encoding, to the sentence embedding of a user’s question can identify source material that may serve as a useful response. And the more training data that you can feed a model, the greater the potential for improving on those results.

Having described what it takes to make a model that’s accurate and precise, Kurpicz-Briki then raises the topic of model bias. What comes out is related to what goes in – it’s how LLMs work. And that can reflect on some undesirable pairs of questions and answers, which can make chatbots generate some ugly and downright dangerous responses.

Mitigating bias is not straightforward. It’s a major undertaking for developers to try and align their model to that they don’t offend users. We learned earlier how it was an advantage to have contextualized word embeddings. But it also means that bias can be highly entangled in LLMs.

Developers may think that they’ve removed everything, but rephrasing a text prompt could elicit previously hidden bias. “Existing detection methods typically identify a very specific bias and mitigate it in the best case, but do not solve the problem as a whole,” the author cautions.

Our latest Vulnerability Rating Taxonomy #VRT release marks a milestone for the #CrowdsourcedSecurity industry.

Are you wondering how? VRT gives customers and #hackers a shared understanding of how #LLM-related vulnerabilities are defined. Learn more: https://t.co/f1WgsJOkZf pic.twitter.com/QA0eFs9z4R

— bugcrowd (@Bugcrowd) December 19, 2023

Having gained a good understanding of how LLMs work, their origins, and some of the dangers that go with the territory, readers are well-placed for the final section of the book, which certainly provides food for thought.

The future of humans and language models is a pressing topic. There are huge advantages in utilizing LLMs to make it possible to query vast amounts of data using natural language prompts. Users can talk to data without having to learn specific commands or spend time formatting information. And that’s just one of the many ways that chatbots can boost productivity for firms.

Don’t overlook the uncanny valley

AI will reshape the business landscape, education, and other key elements of our lives. And having a knowledge of how LLMs work will certainly help in understanding some of those twists and turns.

For example, Kurpicz-Briki raises the issue of the uncanny valley effect, which has its origins in robotics – where designs that are clearly distinguishable from humans are better accepted by users – but also applies to chatbots. It’s a topic that touches on the design of VR applications too.

Users tend to get freaked out when systems mimic humans too closely and billion-parameter statistical text predictors have gotten pretty good at that. It’s another reason to read ‘More than a Chatbot – Language Models Demystified’ – to realize that while LLMs are impressive, they are not magic.

The post Demystifying language models – an explainer on how LLMs work appeared first on TechHQ.

• BrainGPT – a “mind-reading” app – is significantly closer than you might have imagined.
• A significant advance in sensor technology has made BrainGPT a portable prospect.
• At the moment, it’s a technology with around a 40% accuracy – but speech recognition moved quickly, and so could this.

Volunteers used to have to lie inside an fMRI machine and listen to podcasts before large language models (LLMs) could decode their thoughts. But not anymore. Portable, high-performance multichannel wireless EEG acquisition systems have been shown to be capable of BrainGPT-enabled thought-to-text conversion, which is a big deal.

The team behind the latest breakthrough – a group of researchers based at the Human-centric Artificial Intelligence Center in Sydney, Australia – presented their results at NeurIPS 2023, which took place this week in the US.

READ NEXT

Neuromorphic roadmap: are brain-like processors the future of computing?

As a video of the BrainGPT system (embedded in the section below) reveals, the setup connects a so-called multitask EEG encoder with LLMs to decode coherent and readable sentences from EEG signals. In other words, BrainGPT can turn thoughts (measured by wearing a cap of non-invasive electrodes) into words (shown on a screen).

There’s a lot to unpack here, so let’s begin with the EEG acquisition system.

How to design a portable brain scanner

Thought-to-text conversion relies on associating patterns of brain activity with certain words and phrases. And there are a few ways of gathering that information.

To give algorithms the best chance of finding unique patterns that can be tagged to text, high-resolution data is ideal. And the closer that you can place electrodes to the brain neurons themselves, the better. However, there’s also risk and convenience to consider.

It’s notable that the University of Technology Sydney (UTS) team has managed to perform thought-to-text conversion using a non-invasive headset with dry electrodes that sit over the wearer’s hair.

This is the most convenient scenario for users, but unlike brain electrodes used together with a gel or having hardware implanted, the arrangement typically produces the noisiest signals – making it more difficult for AI systems, such as BrainGPT, to read someone’s mind.

The UTS team gets around this in a few ways, starting with materials selection. Its non-invasive sensors for the brain-machine interface are based on micropatterned epitaxial graphene.

Graphene – a one-atom thick form of carbon, first isolated using Scotch tape by Andre Geim and Konstantin Novoselov at the University of Manchester, UK, winning them the 2010 Nobel Prize in Physics – is a so-called ‘wonder material.’

Graphene is incredibly strong, thin, and a great conductor of both electricity and heat. A combination of those properties make the nanomaterial well-suited to measuring cortical electrical activity.

“[Its] very thin nature, combined with high electrical conductivity, biocompatibility, corrosion resistance, and stability in sweat, are all very attractive characteristics for use as sensors for both invasive and noninvasive neural interfaces,” write Chin-Teng Lin – director of the UTS Human-centric AI Center – and colleagues in a related paper.

The epitaxial graphene, which is grown on the tips of micro-patterned silicon structures, was able to detect EEG signals that were comparable (in terms of signal-to-noise ratio) to the gold standard in wet sensors.

Historically, brain-computer interfaces with high signal quality have been heavy and large, but the researchers have shown that it’s possible to build a portable setup.

The wearable EEG measurement device used in their BrainGPT system transmits its 24-bit high-resolution sampled signals, which can be gathered from up to 64 channels, via Wi-Fi and features a Cortex-M4 microcontroller.

How LLMs help to convert thoughts to words

Aside from the cleverly engineered brain-scanning hardware, another key feature is the use of LLMs. The group uses a framework dubbed DeWave that aligns EEG-to-text encodings with pre-trained language models.

Coming soon to a cortex near you. Very, very near you…

It means that while there may still be some uncertainty in the raw measurements themselves, LLMs can highlight the most-likely next word based on the context of the output so far. And this BrainGPT approach helps to push up the overall accuracy of the system.

The team reports that the translation accuracy score is currently around 40% on BLEU-1, which is a scale between zero and one used to compare the similarity of machine-generated results to high-quality references.

Modern speech recognition systems have BLEU-1 scores closer to 90%. And while it’s clear that the BrainGPT system has some way to go, it’s worth recalling where speech recognition was a decade or so ago.

READ NEXT

Brain-inspired computing could solve energy-efficient AI puzzle

Speech recognition systemsback then needed to be trained for hours by each user before they generated sensible results – whereas today, they work well straight out of the box.

Considering the progress being made by a raft of neurotech start-ups, the prospects for thought-to-text conversion are very bright indeed. Already, companies are making use of relatively crude brain wave measurements to inform users about their performance levels and alertness.

List of neurotech start-ups:

• Arctop
• Bitbrain
• BrainQ
• Cerca Magnetics
• EMOTIV
• Inbrain Neuroelectronics
• Muse
• Neurable
• Neuralink
• Neurosity
• Thync

What’s more, brain signals can be converted not just into text, but also into speech and expressions that can be rendered using a digital avatar. And these breakthroughs will help people who would otherwise be unable to communicate.

In previous work, the UTS group has shown how eight non-invasive graphene sensors is sufficient to control a four-legged robot using brain activity alone. And there are lots of inspiring examples of what can be done using consumer hardware that’s starting to become available at relatively affordable prices.

Sour lemons and JavaScript

One of the most curious examples (assuming that the video ‘I literally connected my brain to GPT-4 with JavaScript’ is for real) is being able to trigger applications by simply thinking about sour lemons.

The approach makes use of an SDK available with the Neurosity Crown – a device that claims to increase your productivity by boosting your brain’s concentration.

As well as reporting on the wearer’s focus level, the Crown’s software interface lets users train algorithms to recognize custom thought patterns. Putting the feature to the test, the video’s author imagined biting into a sour lemon, relaxed, and repeated 30 times until the thought pattern had been fully captured.

Using JavaScript, it’s possible to have that data trigger applications by subscribing to those observables and have code run whenever the events are true.

On TechHQ we’ve written previously about the rise of neuromarketing and using brainwaves to capitalize on sales opportunities, and the future could be a strange one indeed. For example, what if your smartphone notified you of nearby sandwich shops when it knew that you were feeling hungry?

Buckle up – brain-reading AI is on the horizon.

The post BrainGPT thoughts to text – how AI can read your mind appeared first on TechHQ.

• FP&A software lets businesses plan for future growth.
• While spreadsheets have their place – and their time – FP&A software allows you to extract more value from your data.
• Generative AI is making its presence felt in FP&A software, bringing new capabilities.

If you are buried in spreadsheets, and the thought of running another quarterly forecast in Excel is making you wince, it could be time to consider financial planning and analysis (FP&A) software. It’s the 21st century, after all.

There’s plenty to celebrate about spreadsheets. Tools such as VisiCalc and Lotus 1-2-3 brought accounting entries made on large sheets of paper into the digital age. Electronic spreadsheets were one of the reasons why many businesses invested in personal computers and opened the door to ‘what if’ scenario planning.

But a lot has happened between then and now, including the growth of FP&A software – dedicated tools that overcome the issues that firms face when trying to chart future growth using Excel or Google Sheets.

Spreadsheet problems

It doesn’t take long to discover the pitfalls of using spreadsheets for business planning. “Excel is not a process-based tool,” Nelson Petracek – chief technology and product officer of Board International, a global provider of intelligent planning solutions – told TechHQ. “At a certain point, firms look for something better.”

READ NEXT

Quantum computing in finance: steampunk chandeliers have their uses

A big disadvantage of using software such as Excel or Google Sheets is that information can quickly become siloed and teams can become isolated in their thinking. Multiple spreadsheets can be in use at the same time, and that creates another problem.

If you’ve ever asked a colleague whether they have the most recent version of a spreadsheet, you’ll be aware of the perils of trying to manage version control manually. On top of that, companies put themselves at risk if sensitive business information is left unprotected.

General spreadsheet applications are not necessarily built with security front and center, and the business risks of using spreadsheets include unauthorized access, suffering a data breach, or losing data if standalone files are deleted by mistake and not backed up.

Last week, a portion of our @BlumbergCapital team met to strategize around fintech.

The most interesting subsectors we identified were:

– insurtech
– fp&a software
– b2b lending
– crypto
– regtech

What’s missing?

— Greg (@gregbrainard) January 30, 2023

It’s no accident that growth in FP&A software has coincided with companies generating more information than ever before. And solutions help not just in remedying the shortcomings of spreadsheets, but also combine other tools for visualization and data management into a single holistic platform.

Difference between business intelligence and FP&A software

Unlike business intelligence solutions – which answer questions such as “What can this data tell me?” – FP&A tools are designed fundamentally to help organizations address planning challenges.

“Companies are trying to expand and want to know what product goes where to meet demand,” said Petracek, referring to an everyday use case for Board’s intelligent planning solution.

Depending on their operating sector, firms may want to map that growth to e-commerce targets and run different scenarios to optimize pricing. Systems can incorporate various influencing factors, such as calendar events, weather forecasts, and many other business-relevant data sources.

List of FP&A software providers –

• Abacum
• Board
• Cube
• Finmark
• Kepion
• Planful
• Pry
• Syntellis
• Telemetry
• Vena
• Workday Adaptive Planning

To get started with FP&A software, users will typically focus on an initial use case, and go through an onboarding process with their provider. Petracek noted that there are often two types of response as users see the solution spring to life.

At the executive level, business leaders appreciate the amount of visibility that FP&A software gives them. Board’s solution includes a ‘Digital Boardroom’ with a built-in briefing book that includes geo, sales, EBITA overviews, and country details.

Annual or quarterly reports can be created within the platform as well as exported as so-called ‘live reporting books’ in MS Word and PowerPoint. What’s more, because those assets are integrated into the intelligent planning solution, charts and in-text references update in tandem with any changes or updates made within the main platform.

Having a single source of truth avoids any confusion between old and new data and means that users can make decisions based on a complete real-time view of business performance.

For staff who’ve been grappling to manually update spreadsheets and paste that information into reports, document automation is a huge time-saver and allows them to perform more value-added tasks.

Plus, as in almost every other enterprise software application available today, the benefits of generative AI and large language models hasn’t escaped business planning tools. The feature lets users have a conversation with their data and quiz the information using natural language, rather than having to learn any specialist commands.

It’s possible to assign access right down to the level of individual cells within the platform, and security around the system – in terms of who can do what – is fully controlled by the customer.

“The cloud environments are completely separate,” said Petracek. “Each customer has their own box.”

Board’s secret sauce includes a proprietary memory technology that it dubs Hybrid Bitwise Memory Pattern (HBMP), which – according to the firm – enables ten times faster query execution.

READ NEXT

Researchers reach for real-time financial predictions with Graphcore silicon

As you would expect, Petracek was reluctant to go into too much detail, but he did reveal that the FP&A software can slice and dice data in different ways so that calculations are available in an instant. Also, the HBMP technology allows Board to scale its solution to match growing data demands.

In terms of deployments, finance teams are often some of the first experience what the software can do. However, as word spreads, organizations will scale up their activity and apply FP&A software across multiple domains – such as manufacturing and other business areas, to address different scenarios.

Having FP&A software is a boost in terms of making it easier for different business units to collaborate. And there are regulatory wins too, as firms can demonstrate that they are on top of their business information and have data to support their claims.

From a usability perspective, web-native FP&A software can be viewed on any device, from a laptop to a smartphone. “People want to consume information in different ways,” Petracek pointed out.

Today, there are thousands of companies that have made the leap from using spreadsheets to dedicated business planning tools. In the case of Board, customers include some very big names such as Coca-Cola, Puma, L’Oréal, KPMG, GSK, BASF, Toyota, KUKA, and Siemens.

Most recently, Euronext – a pan-European stock exchange group with around 1,900 listed issuers and EURO 62 trillion in market capitalization as of Q3 2023 – has chosen to use Board to optimize its financial planning, forecasting, and budgeting.

It speaks to the capability of FP&A software to deliver insights for use cases where spreadsheets would not only struggle, but perhaps fail completely.

The post FP&A software – swapping spreadsheets for SaaS appeared first on TechHQ.

Where and when to use a quantum computer? It’s one of the most common questions that experts, such as Kirk Bresniker – Chief Architect at Hewlett Packard Labs, get asked by business leaders. Enterprises want to know where in the IT portfolio quantum computers will bring the most significant rewards and when is the right time for firms to invest in solutions.

For decades, quantum computing developers have been promising big things from quantum computers, which is understandable. Quantum computers are costly to develop, and being modest about the technology isn’t going to win over investors. However, it’s important to note that quantum computers aren’t universal computing devices.

“Quantum computing promises transformational gains for solving some problems, but little or none for others,” write MIT Sloan School of Management researchers in a paper dubbed “The Quantum Tortoise and the Classical Hare” submitted to arXiv.

Curious about the impact of quantum computing, but don't have a PhD in Physics? Our new article "The Quantum Tortoise and the Classical Hare" may be the framework you need: https://t.co/aOAmObjjtu

— Neil Thompson (@ProfNeilT) November 17, 2023

The team, led by Neil Thompson – whose career includes appointments at Lawrence Livermore National Laboratories, Bain and Company, The United Nations, the World Bank, and the Canadian Parliament – has come up with a simple framework for understanding which problems quantum computing will accelerate (and which it will not).

Quantum computers open the door to probabilistic computing, with quantum gates adding a twist to each of the qubits in the calculation. As the system evolves, the qubits interact and point to the most likely solution to the problem that they’ve been arranged to describe.

READ NEXT

How can quantum computing help my business?

Prodding a bit further, if we consider classical machines as mapping business questions onto maths – a perspective shared by Scott Buchholz, Global Quantum Lead at Deloitte Consulting, at this year’s D-Wave Qubits conference – then quantum computers give us the chance to use physics instead.

It turns out that some questions are easier to map onto physics than others, and this gets to one of the key considerations in the MIT framework on where and when to use a quantum computer.

Much of the talk on progress in quantum computing surrounds the number of qubits. Systems are notoriously noisy, which adds to the number of physical qubits that are required – to facilitate error correction on logical qubits. On top of this, there are multiple ways of engineering the superposition of ones and zeros through the use of superconducting, trapped ion, photonic, or silicon spin qubits.

Each quantum computing developer has its own preferred approach, and as you walk down the path of trying to understand how quantum computing works, the discussion becomes one about the technology. And this is fine. Large companies can engage their R&D teams and have conversations with hardware developers.

When to use a quantum computer – a rule of thumb

However, just as you don’t need to understand what’s happening inside a CPU to benefit from a laptop, companies can focus their attention on the kinds of problems that quantum computers can help with, rather than getting bogged down with the numbers and types of qubits.

In their decision-making framework, Thompson and his colleagues identify two determinants in understanding when to use a quantum computer – the efficiency of the algorithm and the scale of the problem that needs to be solved.

“The problem size matters because the benefit of an algorithmic advantage is larger for larger problems,” explains the team. “This means that if a problem is too small, the classical computer will have already completed the problem by the time the quantum computer’s algorithmic benefit kicks in.”

Quantum computers are often mentioned in terms of being able to tackle problems that are effectively impossible with classical machines. But the researchers want to guide enterprises on other opportunities too, where a quantum economic advantage exists.

Their analysis also considers technology roadmaps so that companies can assess when the window for using a quantum computer could open up for them.

Problems that become exponentially harder to solve as the size of the problem increases are interesting candidates when thinking about alternatives to using classical computing machines. And Thompson and his co-authors – Sukwoong Choi and William Moses – provide a useful rule of thumb.

“If a classical algorithm takes exponential time and there exists a polynomial quantum algorithm, you’re likely to get a speedup,” they comment when discussing their framework on when to use a quantum computer.

Examples of quantum computing as a service (QCaaS) providers

It’s worth pointing out that companies don’t have to invest in bare metal hardware. For most customers, their first experience of what qubits are capable of will be via the cloud using one of a number of QCaaS providers.

Amazon Braket makes it straightforward for firms to work with different types of quantum computers and circuit simulators. Amazon advertises that Braket comes with one free hour of simulation time per month, lowering the cost barrier to getting started.

READ NEXT

Industrial applications of quantum computing add up for D-Wave

QCaaS hardware associated with Bracket includes gate-based superconducting processors from Rigetti and OQC, neutral atom-based quantum processors from QuEra, and IONQ’s gate-based ion-trap processors.

Microsoft’s Azure Quantum cloud service is another option for firms. Here, users get access to systems from Quantinuum, QCI, and PASQAL, as well as the quantum computing hardware mentioned above.

And companies can also access quantum computing solutions in the cloud using QCaaS platforms operated by developers such as IBM, Google, and D-Wave.

There’s no shortage of options, and with frameworks to guide enterprises on where and when to use a quantum computer, now is a good time to think about the types of algorithms supporting your operations and whether qubits can provide an economic advantage to the bottom line.

The post Tips on where and when to use a quantum computer appeared first on TechHQ.