Get SGD58.85 off your premium account! Valid till 9 August 2021. Use the Code ‘SGLEARN2021’ upon checkout. Click Here

Optimising Semiconductor Chip Placement with Machine Learning

Semiconductor chip placement is a critical step in chip design, but it is also a laborious task that can take up to weeks to complete.

FPGA (field-programmable gate array) and ASIC (application-specific integrated circuit) chips, which are widely used in consumer electronics and enterprise systems, can consist of hundreds to millions of transistors depending on their size and complexity. Before they can be manufactured, different modules have to be carefully arranged on the chip.

Given the complexity of today’s semiconductor products, this requires the expertise of highly-skilled and experienced engineers. When carrying out chip placement, the engineers work with rectangular grids, each of which contains parts of one or more modules. Poor chip placement can result in a product that does not meet performance specifications, leading to costly delays in production.

To address these challenges, chip design optimisation company Plunify decided to leverage artificial intelligence (AI) in its quest to help businesses and organisations build better semiconductor products.

Started in 2009 by two passionate engineers, Plunify seeks to improve design performance and save time and resources in the design process.

Working with AI Singapore under the 100 Experiments (100E) programme, it explored the use of machine learning to assist human experts and speed up the chip placement process. Computationally-efficient generative models were used to generate chip placements. Convolutional neural networks and other deep learning architectures were then applied to improve the quality of chip placement by predicting good placements for production.

The outcomes of the project were amazing. A ground-breaking 80 percent accuracy was achieved in applying image recognition to chip placement. This enabled Plunify’s chip design partners to slash the development cycle for products ranging from automotive power chips to 5G communications chips from 2 months to a week.

The prototyping cost for new chips was also reduced by 8 per cent, and new chip designs could be developed 10 times faster.

To find out more about our 100E Programme, please click here.

“The team from AISG provided much-needed expertise and diligence to help us design and craft the data pipeline and models for this effort. This was vital to the success of the project.”
Ng Harn Hua
Co-founder, Plunify

Improving the Accuracy of Spot Price Predictions for Base Metals

The spot prices of base metals, which are used for constructing infrastructure and various types of products, can be incredibly difficult to predict. The prices are stochastic and highly non-linear as they are influenced by many factors with complex relationships. They are also affected by macro-economic situations such as currency exchange rates and government policy changes, such as increase in import tax rates.

Four Elements Capital, a specialised macro and commodity quantitative asset manager, wanted to find out if artificial intelligence (AI) and machine learning (ML) techniques could be used to improve the accuracy of spot price predictions for base metals. Such techniques will leverage not only market data (trading prices and volumes), macroeconomic data, supply and demand data and third-party estimations but also alternative data sources such as news, specialised forums, reports, social media and demand forecasts.

The firm is promoting collaborative research through its Alphien product, which is an open platform to conduct financial research providing the data and the computing power to allow academics to work on financial projects. Its managing director, Lionel Semonin, believes that the ability to integrate AI successfully into the financial domain will position Singapore as the global financial centre of the future.

Taking a step towards this goal, Four Elements worked with AI Singapore through the 100 Experiments (100E) programme to develop an ML framework for price forecasting on the Alphien platform. The framework consisted of three components. The first was the classification of the direction of price movement using an ensemble of ML and deep learning models. The second was a regression of price predictions using a deep learning model. The third was a filter mechanism to align the outputs from the above.

The team also developed new indicators to extract intelligence from alternative data sources relevant to base metal trading, such as news and analyst reports, to derive short-term price movement indicators.

When the source code was deployed on the Alphien server for testing, it was found that the ML framework outperformed traditional linear model benchmarks in its accuracy for out-of-sample forecasts for one, three, five and 10-day horizons.

“There is potential for the ML framework to be spun off and scaled to help financial institutions, business owners and government entities better manage their risk exposure,” said Semonin. “Implementing the AI solution in the financial world will help strengthen and develop Singapore as a leading fintech hub.”

The Learning Never Stops for AIAP Graduate Desiree Chen

When Desiree Chen joined the AI Apprenticeship Programme (AIAP)® in 2020, COVID-19 was already wreaking havoc on the economy. Like many other people, she was afraid of the prospect of unemployment during this tumultuous period. But stronger than that was a desire to acquire a well-rounded set of AI and data science skills.  And so she put aside short-term uncertainties to pursue her long-term plans.

Desiree graduated from Nanyang Technological University in 2008 with an honours degree in Maritime Studies and was working at the Maritime and Port Authority of Singapore for more than seven years when she decided to embark on the Master of Science in Business Analytics programme at the National University of Singapore. The postgraduate studies gave her a good foundation in analytics, and she went on to join Singapore Airlines as a Pricing and Merchandising Analyst.

Her work experiences allowed her to understand how technology can help organisations manage huge amounts of data in a more efficient manner, and she often found herself at the intersection of business and technology. “That allowed me to grow professionally, to empathise and work better with business units and technology teams,” she said.

By the time Desiree applied for AIAP, she had already built up a good working knowledge of AI, both on her own and through her work.

“I spent about five years prior to AIAP building my foundation in Machine Learning (ML) and gaining work experience in various aspects of AI and data science,” she said.

However, she felt that her knowledge was incomplete and wanted to deepen her skills. “I wanted to be exposed more to the engineering aspects of AI and data science projects,” she said.  “In my interactions with industry, I observed this growing and crucial need to have a good infrastructure in place, so that organisations can really grow their AI and data science capabilities and not run ML models locally on laptops.”

She knew that AIAP would provide her with the opportunity to hone her engineering skills and equip her with the knowledge and hands-on experience to deal better with these areas of AI.

For Desiree, the most interesting aspect of the apprenticeship programme was the opportunity to work on a real-life data engineering project – something which she did not get to focus on in her previous roles.

During the seven-month project phase of AIAP, her team was involved in integrating data versioning and data cataloging features into a data platform to support industry projects undertaken by AI Singapore.

As someone who was relatively unfamiliar with software engineering, one of Desiree’s concerns before she joined AIAP was whether she could keep up with the rest of her cohort.

She needn’t have worried. Whenever she encountered technology stacks that she was not very familiar with, she found that she could always approach her mentors and team mates for technical advice and guidance.

In fact, one of the things that Desiree really liked about AIAP was the culture of knowledge sharing. “Those around me were helpful and forthcoming in sharing their expertise, and the mentors always encouraged us to ask questions. I learned to ask the right questions to get the answers that I needed,” she said.

Through regular architecture design meetings, Desiree also developed a better understanding of the various trade-offs in architecture design. And she felt a great sense of achievement when she was able to integrate data versioning into the data platform and demonstrate to AI engineers and fellow apprentices how the feature could be used in the various projects undertaken by AI Singapore.

While all this was ongoing, Desiree also managed to squeeze in the time to obtain her AWS Certified Solutions Architect – Associate and AWS Certified Machine Learning – Specialty certifications.

“Technology is ever evolving, so the learning never stops”

Desiree graduated from AIAP on 9 April 2021 and joined Singapore Exchange Limited as Vice President of the CFO Unit (Data and Business Analytics). Under the guidance of AI Singapore’s Talent & Career Management Team which provides career advisory and development support to the apprentices, she secured this role and gets to apply the data engineering skills gained from AIAP and her analytics expertise from previous work experiences.

Desiree sees AIAP as an important step in her growth as an AI and data science practitioner. To those contemplating the same path, she has this to share: “Do not choose this journey just because everyone says that it is the next big thing or that you can earn a higher salary. While that may be the case, you have to ask yourself if you are truly passionate about this. You need to have passion for those parts of the journey where it gets tough, when you get tired and you doubt yourself. It is not all smooth-sailing. You need to be resilient.”

“This is a journey in your career,” she said. “Think of it more as a marathon than a sprint to the finish line, and learn to enjoy the ups and downs.”

uParcel Boosts Delivery Efficiency by 20% with AI

It’s no longer fastest fingers first. With a new clustering algorithm, uParcel, an on demand 24/7 courier service in Singapore, has found a way to allocate jobs more fairly and efficiently to drivers while ensuring faster deliveries for customers.

uParcel is the largest home-grown same-day delivery company with an on-demand platform that supports a decentralised distributed model. This means that the company delivers packages point to point without consolidating them at warehouses. Instead, the uParcel mobile app matches delivery bookings to their network of crowdsourced drivers in real time, providing fast and flexible delivery options for the growing ecommerce sector.

Without the use of artificial intelligence (AI), the process for this involves posting the delivery jobs online. The drivers then choose the ones that they want to take on.

The disadvantage of this arrangement is that drivers have to manually search through the list of delivery orders on their app, and those with the fastest fingers will clinch the coveted jobs. This offers a less-than-ideal user experience and also means that multiple drivers could be taking different jobs with similar pickup and delivery points. The system was thus inefficient, with each driver operating way below his or her optimal capacity.  

To boost driver efficiency, uParcel wanted to find a way to bundle jobs so that drivers can take on more jobs with similar pickup and delivery points and maximise their carrying capacity. But with its internal engineering team already stretched supporting its daily operations, it needed external help.

uParcel reached out to AI Singapore (AISG) to find out how it could tap on AI to achieve its goal. Working with the company in just 20-man days, the AISG team developed a clustering algorithm that groups pickup and destination points based on their proximity to each other. The algorithm is applied on a set of orders that contains the time and location of each pickup or delivery. 

"Out of hundreds of collection points, we are able to identify collection points of different merchants that are close together, and also group deliveries points based on their distance from each other instead of being restricted by rigid postal regions. This allows us to achieve much greater flexibility in job allocation,”
William Ng
Chief Operations Officer, uParcel

The optimised batching of pickups based on location has improved driver productivity by up to 20 percent.

Previously, there can be multiple collection points for drivers but they will have to manually consider if they are near each other and search through the jobs. With the new batching algorithm, it presents to them an optimally grouped batch of jobs that saves drivers precious time in considering the options. Multiple collection points within a 2km radius can be batched, enabling drivers to complete more jobs in a single trip. This increases their efficiency and income and at the same time, customer service receives a boost with faster pickup and delivery. 

The AI solution, which is currently in beta testing stage, is set for a full launch in June.

“The collaboration with AISG has enabled us to turn data into actionable insights in record time,” said William.

mailing list sign up

Mailing List Sign Up C360