Large language models (LLMs) have shown exceptional capabilities in comprehending human language, reasoning, and knowledge acquisition, suggesting their potential to serve as autonomous agents. However, training high-performance web agents based on open LLMs within online environments, such as WebArena, faces several critical challenges. The first challenge is insufficient predefined training tasks in online benchmarks. The next challenge is assessing success for arbitrary web browsing tasks due to the sparsity and high cost of feedback signals. Lastly, the absence of a predefined training set necessitates online exploration, leading to policy distribution drift and potential catastrophic forgetting, which can decrease the agent’s performance over time.

The existing methods include adopting LLMs as Agents and Reinforcement Learning (RL) for LLMs. Current research in LLMs as Agents has two main categories: training-free and training-based approaches. While some studies have used powerful LLMs like GPT-4 to generate demonstrations, the accuracy of these methods remains insufficient for complex tasks. Researchers have explored RL techniques to address this challenge, which uses sequential decision-making to control devices and interact with complex environments. Existing RL-based methods, such as AgentQ, which uses DPO for policy updates, and actor-critic architectures, have shown promise in complex device control tasks. However, the limited and sparse feedback signals are often binary success or failure after multiple interaction rounds in web-based tasks.

Researchers from Tsinghua University and Zhipu AI have proposed WEBRL, a self-evolving online curriculum RL framework designed to train high-performance web agents using open LLMs. It addresses the key challenges in building LLM web agents, including the scarcity of training tasks, sparse feedback signals, and policy distribution drift in online learning. Moreover, it utilizes three key components: 

• A self-evolving curriculum that generates new tasks from unsuccessful attempts.
• A robust outcome-supervised reward model (ORM)
• Adaptive RL strategies to ensure consistent improvements. 

Moreover, WEBRL bridges the gap between open and proprietary LLM-based web agents, creating a way for more accessible and powerful autonomous web interaction systems.

WEBRL utilizes a self-evolving online curriculum that harnesses the trial-and-error process inherent in exploration to address the scarcity of web agent training tasks. In each training phase, WEBRL autonomously generates novel tasks from unsuccessful attempts in the preceding phase, providing a progressive learning trajectory. It also incorporates a KL-divergence term between the reference and actor policies into its learning algorithm to reduce the policy distribution shift induced by curriculum-based RL. This constraint on policy updates promotes stability and prevents catastrophic forgetting. Moreover, WEBRL implements an experience replay buffer augmented with a novel actor confidence filtering strategy.

The results obtained for Llama-3.1-8B trained using WEBRL achieve an average accuracy of 42.4%, surpassing all the baseline approaches, including prompting and training alternatives. WEBRL excels in specific tasks such as Gitlab (46.7%) and CMS (54.3%), showcasing its ability to address complex web tasks effectively. Moreover, it outperforms imitation learning-based methods, such as SFT and Filtered BC. Moreover, it consistently outperforms DigiRL, a previous state-of-the-art method that conducts policy updates on a predefined, fixed set of tasks, which may not align with the model’s current skill level. WEBRL addresses this by using self-evolving curriculum learning, adjusting the task complexity based on the model’s abilities, promoting wider exploration, and supporting continuous improvement.

In this paper, the researchers have introduced WEBRL, a novel self-evolving online curriculum RL framework for training LLM-based web agents. It addresses the critical challenges in building effective LLM web agents, including the scarcity of training tasks, the sparsity of feedback signals, and the policy distribution drift in online learning. The results demonstrate that WEBRL enables LLM-based web agents to outperform existing state-of-the-art approaches, including proprietary LLM APIs, and these findings help enhance the capabilities of open-source LLMs for web-based tasks, paving the way for more accessible and powerful autonomous web interaction systems. The successful application of WEBRL across different LLM architectures, like Llama-3.1 and GLM-4 validates the robustness and adaptability of the proposed framework.

Check out the Paper. All credit for this research goes to the researchers of this project. Also, don’t forget to follow us on Twitter and join our Telegram Channel and LinkedIn Group. If you like our work, you will love our newsletter.. Don’t Forget to join our 55k+ ML SubReddit.

[Sponsorship Opportunity with us] Promote Your Research/Product/Webinar with 1Million+ Monthly Readers and 500k+ Community Members

The post WEBRL: A Self-Evolving Online Curriculum Reinforcement Learning Framework for Training High-Performance Web Agents with Open LLMs appeared first on MarkTechPost.

The post WEBRL: A Self-Evolving Online Curriculum Reinforcement Learning Framework for Training High-Performance Web Agents with Open LLMs appeared first on BusinessViewed.

Lynn Hazan, founder of ChicpeaJC, left a corporate job to pursue her creative passions, turning a blog into a successful social media agency. Her journey from the corporate world to entrepreneurship reveals the power of self-belief, adaptability, and purpose. Here, Lynn shares her experiences, challenges, and advice for other moms who want to build something meaningful.

1. Lynn, what inspired you to leave corporate life and start ChicpeaJC?

I had been working in corporate for about seven years, but it never felt like “me.” Both of my parents are entrepreneurs, so I grew up with that spirit, but here I was, working a job that didn’t allow me to tap into my creativity. I was also a new mom, so I couldn’t just quit my job, but I knew I wanted more. That’s when I decided to start a blog. I spent six months creating content and building the website, and when it launched, it took off! Two months in, I figured out how to monetize it, and eventually, I was able to leave my job. A few years later, I saw a growing need for social media content among local businesses, so I started managing social media for a coffee shop as a test. Now, 10 years later, I manage social media for over 15 businesses.

2. What challenges have you faced in growing your business?

There have been a lot! I’ve dealt with issues like employee challenges, losing my office due to high rents, and of course, the impact of COVID. I lost most of my clients during the pandemic and had to rebuild from scratch. Balancing work with motherhood, especially without maternity leave, was tough, too. But having a supportive partner, believing in myself, and taking it one day at a time helped me push through. Each obstacle strengthened my resilience and faith in my journey.

3. What keeps you motivated, even in difficult times?

Knowing that this business supports my family motivates me. I don’t have the luxury of giving up! I also believe in the universe and that everything happens for a reason. Whenever things feel overwhelming, I remind myself that setbacks often bring valuable lessons. What doesn’t kill you truly makes you stronger.

4. What advice would you give to someone thinking about starting their own business?

Start now! There’s never a “perfect” time, and yes, there will always be someone who seems to be doing it better. But there’s only one you, and that’s your superpower. Success is about persistence—every time I faced a setback, I just got back up. Success isn’t defined by avoiding failures; it’s defined by how often you get back up. And remember to enjoy the process, not just the destination.

5. What has been a defining moment in your entrepreneurial journey?

For me, it’s been creating something from scratch that combines my passions and supports my family. Building a business around what I love and seeing it grow into something that can make a real difference in people’s lives is incredibly fulfilling. It’s a reminder that we all have the power to create something meaningful.

6. What’s your vision for the future?

I joke with my husband that I’d love to be a stay-at-home Pilates wife someday! But honestly, I know I’ll always want to keep evolving. I’m really passionate about finding ways to give back to the Jewish community and making a positive impact in new ways. I want my work to continue to align with my values.

7. What message would you like to share with other moms?

I want to encourage moms to pursue what they truly want—whether it’s a career, staying home with their kids, or doing both. You don’t have to compromise on your dreams. I hope my journey shows that it’s possible to create a life you love while staying true to yourself. There’s no one right way to be a mom or an entrepreneur, and it’s okay to carve out your own path.

Final Thoughts

Lynn Hazan’s story shows that entrepreneurship doesn’t require a perfect plan; it requires passion, resilience, and self-belief. She proves that you can build a meaningful business while honoring your values and supporting your family.

To learn more about Lynn and her work, visit chicpeajc.com.

The post Lynn Hazan on Transforming a Simple Blog into a Thriving Business: Inspiring Moms to Chase Their Dreams appeared first on BusinessViewed.

Colden Blades, founder of Blades Marketing, started his entrepreneurial journey during his freshman year at SUNY Plattsburgh. What began as a campus t-shirt business evolved into a full-scale digital marketing agency, now in operation for over five years. In this interview, Colden shares insights into the lessons learned, the importance of purpose over happiness, and his vision of living a life rooted in principles.

1. Colden, what sparked your journey into entrepreneurship, and how did your first business come about?

It all started when I joined clubs, organizations, and fraternity life on campus at SUNY Plattsburgh. I noticed that every semester, these groups were buying promotional clothing items, which gave me the idea to start a t-shirt company. I pitched the idea to my good friend Frank Tricamo during finals week, and he was on board immediately. After Christmas Break, we jumped in and started the business. That first experience taught me so much about running a business, which translated directly into launching my digital marketing company two years later. I’m forever grateful for that partnership and the lesson that surrounding yourself with ambitious people is essential.

2. What has been one of the biggest challenges in your entrepreneurial journey, and how did you navigate it?

The greatest challenge was realizing that, in the beginning, the goal shouldn’t be to build a successful business but to build yourself into someone capable of building a successful business. I learned this lesson the hard way by trying to scale before I was ready—like building a tower on a cracked foundation. It’s only a matter of time before everything comes crashing down if you don’t have the personal growth to support it. That realization made me focus more on developing myself before scaling the business.

3. You mentioned a mental shift that helped you during hard times. Can you tell us more about that?

Absolutely. For a long time, I treated happiness as my North Star. But happiness, like motivation, ebbs and flows. The shift came when I replaced happiness with purpose. I realized that, often, the things that help us grow involve hardship, suffering, or discomfort. When purpose became my guiding force, it didn’t matter if a task made me happy or if I felt motivated. If I believed it would bring me closer to my purpose, I got it done. This shift enabled me to stay focused and push through the difficult moments.

4. What advice would you give to other entrepreneurs about building a successful business?

Surround yourself with people who challenge you to grow. You become the average of those you spend the most time with. Also, focus on becoming the person capable of building and sustaining success rather than just the business itself. And lastly, have a guiding purpose. It will sustain you through times when motivation and happiness might be lacking.

5. What has been your proudest moment so far in your entrepreneurial journey?

My proudest moment has been going from barely being able to keep my own head above water to being able to take care of the people I love. It’s incredibly rewarding to be in a position where I can give back to the people who supported me along the way.

6. What’s your vision for the future of Blades Marketing?

My vision is simple: to remain true to my morals, values, and principles in everything I do. I want to live a life of purpose and die with my boots on, as they say. I hope to show others that you can lead an ambitious life of purpose without compromising your values along the way.

7. What impact do you hope to make through your work and example?

I genuinely hope to set an example for other men that you can pursue a life of purpose and ambition without sacrificing your morals, values, or principles. In a world where it’s easy to compromise, I want to demonstrate that success and integrity can go hand in hand.

Final Thoughts

Colden Blades’ journey from college entrepreneur to digital marketing agency owner highlights the importance of building oneself alongside one’s business. His commitment to purpose and integrity offers valuable lessons for aspiring entrepreneurs.

To learn more about Colden and his work, visit bladesmarketing.com.

The post From Campus T-Shirts to Digital Marketing: Colden Blades on Building a Purpose-Driven Business appeared first on BusinessViewed.

1. Phil, you started in the HVAC industry. What inspired you to transition from hands-on work to digital marketing?

I started my career in HVAC, which was comfortable for me since I enjoyed the technical challenges. But after becoming an HVAC business owner, I quickly realized that delivering quality service was only part of the equation. Success also depended on consistently finding clients and generating leads, and that’s where the challenge lay. Marketing was a world I knew little about, but I dove into it because I knew it was essential for my business to thrive. Over time, learning SEO, web development, and lead generation became my new focus.

2. What were some of the biggest challenges you faced when learning digital marketing?

The learning curve was intense. I spent countless hours studying, watching tutorials, and experimenting with digital tools. There were times when technical issues or complex concepts like SEO algorithms left me feeling overwhelmed. One night, I was watching a webinar about domain registrars, hosting, and DNS settings, and by the end, I felt completely lost. It was my wife’s support that kept me going. Her unwavering belief in me reminded me that I’d overcome challenges before, and I could do it again. Enrolling in a top marketing and SEO training course was also a turning point, as it gave me real-world insights and confidence.

3. You’ve mentioned that your wife’s support was crucial during tough times. How did her encouragement impact your journey?

My wife was—and still is—my rock. During moments of self-doubt, her calm reassurance reminded me of my capabilities and determination. There were days when I questioned if I could truly master everything needed to run a successful marketing business, but her belief in me never wavered. Her support was what helped me persevere through the frustrations and push through moments when giving up felt easier.

4. What advice would you give to someone transitioning into a new field or starting a business from scratch?

Embrace the uncertainty and be prepared to fail. I had to navigate a steep learning curve, especially moving from HVAC work to marketing. The learning curve was filled with doubt, confusion, and self-doubt, but I wish someone had told me that those feelings are normal. Also, having a strong support system is essential—people who believe in your vision will lift you up when you’re down. Lastly, keep learning and adapting. You don’t need to know everything right away, but your willingness to grow and learn will set you apart.

5. What was a proud moment for you in building Skill Trade Marketing?

One of my proudest moments was signing my first client. It was a validation of all the hard work and learning I had invested in the agency. But the real pride came when I was able to rank them on page 1 of Google for the top keywords in their niche. Seeing them achieve their growth goals was incredibly rewarding, as I knew my efforts directly contributed to their success. That first win gave me the confidence to keep pushing forward.

6. What’s your vision for the future of Skill Trade Marketing?

I want to continue scaling my agency to help more businesses, especially in trades like plumbing, HVAC, electrical, landscaping, and commercial cleaning, reach their full potential. Our current niche website is Skill Trade Marketing, and I’m focused on refining our strategies to keep our clients ahead in their industries. Beyond that, I aim to become a trusted resource for business owners—not only in delivering results but also in educating them on sustainable growth. My goal is to create lasting partnerships where we grow together year after year.

7. What impact do you hope to make in the lives of your clients?

I want to empower business owners, especially in niche industries, with the tools and knowledge they need to succeed. My hope is to transform how small businesses approach their online presence, showing them that with the right strategy, they can compete and thrive. My clients’ success is my success, and I’m committed to helping them achieve long-term growth and sustainability.

Final Thoughts

Phil Furlong’s journey shows that even the steepest learning curves can be overcome with perseverance, support, and a commitment to growth. From the HVAC industry to SEO, Phil has turned his experience and insights into a thriving agency dedicated to helping niche businesses grow online.

To learn more about Phil’s work, visit furlongseomarketing.com.

The post From HVAC Employee to SEO Expert: How Phil Furlong Built a Thriving Marketing Agency appeared first on BusinessViewed.

In the ever-evolving world of entrepreneurship, adaptability isn’t just a skill—it’s a necessity. Today, we speak with Lauren Stanko, founder of Mojo Media Solutions, a brand development agency that’s turning visions into reality. From navigating the uncertainties of starting a business without a roadmap to launching her passion project, OOMA Mushrooms, Lauren shares her journey of resilience, growth, and the importance of soft skills in moving the needle.

1. Lauren, you’re entering your second year as a full-time business owner. Can you tell us about your journey and what inspired you to start Mojo Media Solutions?

I own a brand development agency that helps take entire visions and bring them to life. When I started, there was no framework to follow—just my soft skills guiding the way. Our most recent project, OOMA Mushrooms, has recently launched, and it’s been incredible to see how the business has taken shape in wonderful ways that weren’t on the original draft. I was inspired to create systems that bring value into the world and offer people alternative solutions to modern mental health problems.

2. What were some of the biggest challenges you faced when starting your business, and how did you overcome them?

One of the hardest parts was never knowing what to do next. Without a framework, I had to build a plan, test it, and quickly adapt if things weren’t going in the right direction. Losing my first client was a significant hurdle. It was a mutual parting—they needed skills that weren’t my strengths. Rather than trying to mold myself into areas outside my expertise, I communicated that a better fit was out there for them. It was hard on my ego, but ultimately, it allowed me to get clear on my strengths and move in a better direction for success.

3. How did that experience shape your approach to business moving forward?

It taught me the importance of adaptability. You can teach yourself almost any hard skill—coding, marketing, business management—but mastering adaptability, resilience, and motivation is much harder. Those are the real skills that move the needle and encourage staying power. The experience helped me focus on my strengths and be more intentional about the clients I work with.

4. What keeps you motivated during tough times or when facing uncertainty?

I remind myself of the remarkable moments, like when we got our first order in the first brand I created. That was an incredible feeling, especially knowing that only about 10% of people who attempt to sell online make sales. Seeing tangible results from hard work fuels my motivation. Also, the vision of OOMA Mushrooms making it into the lives of hundreds of thousands of people struggling with their mental health keeps me pushing forward.

5. What advice would you give to new entrepreneurs just starting out?

You have to develop adaptability. While hard skills are important, adaptability, resiliency, and motivation are the real drivers of success. They encourage staying power. Don’t be afraid to test your plans and quickly adapt if things aren’t going in the right direction. Also, be honest about your strengths and don’t try to be everything to everyone.

6. What’s your vision for the future of Mojo Media Solutions and OOMA Mushrooms?

OOMA is a true passion project. I envision it reaching hundreds of thousands of people who are struggling with their mental health or simply looking to improve their lives. For Mojo Media Solutions, I want to continue creating systems that bring value into the world, helping to turn more visions into reality.

7. What impact do you hope to make through your work?

I aim to offer people alternative solutions to modern mental health problems. By bringing projects like OOMA Mushrooms to life, I hope to make a positive impact on people’s well-being. Through my agency, I want to help others realize their visions, creating a ripple effect of positive change.

Final Thoughts

Lauren Stanko’s journey underscores the power of adaptability and the importance of focusing on one’s strengths. Her commitment to offering alternative solutions for mental health and helping others bring their visions to life is a testament to what can be achieved when resilience meets passion.

To learn more about Lauren’s work, visit mojomediasolutions.io.

The post You Have to Develop Adaptability: How Losing Her First Client Led Lauren Stanko to Success appeared first on BusinessViewed.

1. Vince, what led you to start selling digital products, and how was the transition from software development?

I started about a year ago. Coming from a software development and web agency background, I didn’t know much about digital products or digital marketing. The transition wasn’t easy because I was stepping into a completely different space with a lot of noise. There were tons of so-called “experts,” especially on Instagram, but I had no idea how the average person could actually start making money with digital products. It was overwhelming to figure out who to follow and where to begin.

2. What were some of the biggest challenges you faced in the early stages?

The toughest part was the self-doubt and inconsistency, especially when I wasn’t making any money. There were times when I really questioned if this path was for me, and those doubts would creep in often. During those low points, I just had to push through, stay consistent, and keep going. It was all about moving past that plateau and trusting the process.

3. What kept you motivated to continue, even during those tough times?

One of the biggest motivators was seeing small wins along the way and knowing that progress was possible. Selling my first product for over $1,000 was a huge accomplishment for me. It took time to get there, but that sale helped me believe in myself and showed me that I was on the right path. Staying consistent and building on those small victories kept me going.

4. What advice would you give to someone just starting out in the digital products space?

Find someone you can follow and trust. There’s so much information out there, and it’s easy to get lost. Confidence is key—it’s not just about mindset; it’s about believing in yourself and building that confidence over time. And it’s important not to lose that confidence when things get tough, which happens to all of us. Consistency and confidence go hand-in-hand in building a successful business.

5. What’s been your proudest moment so far in this journey?

Selling my first high-ticket product for over $1,000 was definitely a proud moment. It represented a lot of work and overcoming self-doubt. That sale felt like a milestone, and it reassured me that I was moving in the right direction.

6. What’s your vision for MRR Products Vault and your future in digital products?

My vision is to provide accessible digital products for people just starting their journey. I want to help others jump-start their path to making money online, especially those who want to work for themselves and gain financial independence. There’s something incredibly rewarding about helping others realize that they, too, can achieve success in this space.

Final Thoughts

Vince Edwards’ journey highlights the importance of consistency, confidence, and finding trusted guidance in the digital product space. His story is a testament to overcoming self-doubt and staying focused on small wins that add up over time.

To learn more about Vince and MRR Products Vault, visit mrrproductsvault.com.

The post From Web Developer to Digital Product Entrepreneur: Vince Edwards on Finding Confidence and Consistency appeared first on BusinessViewed.

Today, we’re diving into the entrepreneurial journey of Tyler Cook, founder of HyperMedia Marketing. Starting in 2019, Tyler left his previous agency role to offer a full suite of marketing services. However, it didn’t take long for him to realize that trying to do everything was holding him back. In this interview, Tyler shares hard-won lessons on niching down, building a sustainable business, and the power of consistency.

1. Tyler, what made you decide to leave your previous role and start HyperMedia Marketing?

In 2019, I was working at a small agency that primarily focused on marketing automation. I quickly realized that clients needed more—a holistic approach that included everything from copywriting and paid ads to conversion rate optimization. I felt like I was handcuffed by the limited services we offered, so I decided to go out on my own and launch HyperMedia Marketing, where I could expand on those services.

2. What were some of the biggest challenges you faced in the early days?

One of the toughest challenges was getting clients, which is still a challenge sometimes. The agency space is saturated, and finding a consistent growth channel took time. Another big hurdle was the services-versus-pricing dilemma. When I first started, I was a full-funnel agency, offering everything from ads to website optimization and social media management. I was doing it all, but I wasn’t charging enough to afford any help, which led to burnout pretty quickly. I eventually realized that I loved email marketing, list management, and sales pipeline automation, so I decided to niche down and focus on those areas. This shift helped me attract clients willing to invest in premium services and allowed me to hire help, freeing up my time for growth.

3. You mentioned that niching down was a turning point. How did that decision impact your business?

Niching down was a game-changer. When I scaled back to just email marketing, list management, and sales pipeline automation, I felt a huge weight lift off my shoulders. It allowed me to refine my services, build systems, and create processes that made everything easier. I could also focus on finding clients who understood the value of specialized services. It was a relief to narrow down who I worked with and become world-class in a specific area. That focus made it easier to communicate my expertise and build a sustainable business.

4. What keeps you motivated, especially during the tough times?

When things get difficult, I remind myself of the good months and how small, consistent actions lead to long-term success. If I ever notice a dip in clients or my pipeline, it usually comes down to not doing the small daily tasks that build momentum. Remembering that progress is gradual and rooted in consistency has been crucial for me.

5. What advice would you give to entrepreneurs who are just starting out?

I’d say to anyone starting out: be prepared for the challenges. Running an agency can be harder, more stressful, and lonelier than you might expect, but it’s worth it. You’ll learn valuable lessons and experience incredible wins along the way. Another piece of advice is to focus on personal growth as much as business growth. If you’re not learning and adapting, the business can quickly consume you. And, of course, find your niche. Going deep into a specific area you love will allow you to attract clients who value your work and are willing to invest in what you offer.

6. What’s your vision for the future of HyperMedia Marketing?

My vision is to continue offering niche services at an exceptionally high level, helping businesses succeed through email marketing and relationship building with their subscribers. I want to create thriving relationships between businesses and their clients, which ultimately leads to long-term success for everyone involved.

Final Thoughts

Tyler Cook’s journey highlights the importance of finding your niche, building consistency, and focusing on quality over quantity. His story is a reminder that sometimes, scaling back and specializing can be the key to sustainable growth.

To learn more about Tyler and HyperMedia Marketing, visit hypermediamarketing.net.

The post 3 Lessons I Learned the Hard Way About Running a Small Digital Agency: An Interview with Tyler Cook appeared first on BusinessViewed.

Forefront RF, a Cambridge-based fabless semiconductor startup, has snapped £16 million in Series A funding. The round saw participation from existing investors BGF (which recently invested in Twenty7tec and Kubus) and Foresight Group, alongside new VC partners, Octopus Ventures (which recently backed Semble and Cambri) and Cambridge Innovation Capital.

This investment follows the £3.75 million funding raised last year. The company intends to use the funds to expand operations and its development efforts. The company is set to transition from an early-stage startup to a mature player in RF technology, gearing up for the 2026 launch of its first-generation product, the FFM51010, to the fast-growing cellular-enabled smartwatch market.

Next-gen connectivity solutions 

Founded by Dr. Leo Laughlin and Julian Hildersley in 2020, Forefront RF aims to make multi-band smartphones, wearables, and IoT Devices simpler to design and more globally accessible. It empowers mobile device manufacturers to overcome printed circuit board (PCB) size limitations while supporting more frequency bands with simplified, cost-effective designs. 

The company’s first solution is aimed cellular-enabled wearables, where every millimetre of PCB is critical and expanding frequency band capabilities is essential. By replacing bulky fixed-frequency filters and duplexers with a self-configuring tunable duplexer, Forefront RF enables devices to dynamically adjust to any required frequency, reducing overall component list and manufacturing complexity. This innovation not only saves valuable PCB space but also minimises variants, reducing waste and enhancing supply chain efficiency. 

Since its debut, the company has pushed the boundaries of wireless technology with its patented Foretune technology. It is a transformative self-interference cancellation system that addresses leakage and interference issues inherent in all wireless devices. By inverting and neutralising unwanted signals similar to noise-cancelling headphones, Foretune enables cleaner signal transmission and reception in software-defined radio systems. This helps simplify product design and production processes. With Foretune, Forefront RF is poised to redefine RF standards and drive significant advancements in next-generation wireless technology.

Ronald Wilting, CEO of Forefront RF said: “Securing this Series A funding underscores the strong confidence our investors have in Forefront RF’s vision and market potential. With continued support from Foresight Group and BGF, and new partnerships with Octopus Ventures and Cambridge Innovation Capital, we’re well-positioned to scale rapidly from engineering prototypes to mass production.” 

Investors’ views 

“Forefront RF has created a solution to one of the greatest bottlenecks in mobile data connectivity, based on a decade of advanced research and development. Cambridge Innovation Capital is delighted to partner with Forefront RF as it prepares to launch its highly differentiated product lines into the growing wearables market. We look forward to supporting the team as they scale to deliver global impact across billions of connected devices,” said Edward Inns of Cambridge Innovation Capital.

Owen Metters, investor at Octopus Ventures, continued; “We’re delighted to be supporting Forefront RF on their journey, and are really excited by the Team’s potential. Their patented technology will revolutionise how mobile devices are designed, reducing complexity, and streamlining supply chains. The continuing proliferation of cellular-enabled devices means there is a significant opportunity for technology such as ForetuneTM.”

Andy Bloxam, Head of FTW at Foresight stated: “We are excited to be making this follow-on investment into Forefront RF, a developer of breakthrough tuneable duplexers that can potentially underpin next-generation RF-enabled devices. This is a testament to Foresight’s long-standing commitment to the Cambridge ecosystem, which is widely recognised as a great place to scale disruptive deep-tech businesses such as Forefront.”

Luke Rajah, investor at BGF, iterated: “Forefront RF’s breakthrough technology addresses key technological challenges in an ever-more connected world, while also delivering environmental benefits. Supporting businesses as they transition from start-up to maturity is a fundamental part of BGF’s mission, and it’s immensely rewarding to back such an innovative British business as it develops world-leading technology.”

The post ForeFront RF snaps £16M to make smartphones, IoT devices more accessible appeared first on Tech Funding News.

The post ForeFront RF snaps £16M to make smartphones, IoT devices more accessible appeared first on BusinessViewed.

Machine learning research has advanced toward models that can autonomously design and discover data structures tailored to specific computational tasks, such as nearest neighbor (NN) search. This shift in methodology allows models to learn not only the structure of data but also how to optimize query responses, minimizing storage needs and computation time. Machine learning now moves beyond traditional data processing, tackling structural optimization in data to create adaptable frameworks that exploit distribution patterns and data characteristics. Such adaptability is valuable across fields where efficient data retrieval is crucial, especially in domains where speed and storage are constrained.

Designing efficient data structures remains a significant challenge. Existing structures, like binary search trees and k-d trees, are typically designed with worst-case scenarios in mind. While this ensures reliable performance, it also means they do not leverage potential patterns in the data for more efficient querying. Consequently, many traditional data structures cannot capitalize on characteristics unique to each dataset, resulting in suboptimal performance for queries that could otherwise benefit from custom, adaptive structures. As a result, there is growing interest in data structures that can adapt to specific data distributions, offering faster query times and reduced memory usage tailored to particular applications.

Methods developed to improve data structure efficiency have primarily focused on learning-augmented algorithms, where traditional data structures are modified with machine learning predictions to speed up queries. However, even these methods are limited by their reliance on predefined structures that may need to be optimally suited to the dataset. For example, while learning-augmented trees and locality-sensitive hashing improve search efficiency by combining algorithmic principles with predictive models, they are constrained by human-defined structures. These models still depend on initial data structures, limiting their ability to adapt to unique data distributions autonomously.

The researchers from Universite de Montreal Mila, HEC Montreal Mila, Microsoft Research, University of Southern California, and Stanford University proposed an innovative framework that leverages machine learning to discover data structures suited for specific tasks autonomously. This framework includes two core components: 

1. A data-processing network that arranges raw data into optimized structures
2. A query-execution network that efficiently navigates the structured data for retrieval

Both networks undergo joint end-to-end training, allowing them to adapt to varying data distributions. By eliminating the need for predefined structures, the framework autonomously designs optimized configurations that outperform traditional methods across different data and query types, including NN search and frequency estimation in streaming data.

The methodology involves an 8-layer transformer model, where the data-processing network ranks elements within a dataset, organizing them into efficient configurations. This ranking is refined through a differentiable sorting function, which orders the data based on their ranks. Meanwhile, the query-execution network, comprising multiple independent models, learns an optimal search strategy for retrieving specific data points based on historical query patterns. This joint training customizes the data structure and enhances query accuracy. For instance, the model demonstrates high precision in correctly ordering 99.5% of 1D NN search data despite needing to be explicitly programmed. This level of precision exemplifies how data-driven structures, once designed, can enhance both storage efficiency and retrieval speed.

The framework excelled across several test scenarios in evaluations. In 1D NN search, the model displayed accuracy levels higher than traditional binary search methods. For example, the model outperformed binary search by strategically initiating queries closer to the target location when tested on data with a uniform distribution over (-1, 1) with 100 elements and limited to seven lookups. In high-dimensional contexts, such as 30-dimensional hyperspheres, the model used projections that closely resembled locality-sensitive hashing, achieving results comparable to specialized algorithms. Notably, in a challenging setup where query accuracy must be completed within a limited space, the model utilized additional space effectively, trading memory for query precision. The model’s accuracy increased when given seven extra vectors for storage, demonstrating its adaptability to varying spatial constraints.

The research presents several key takeaways that illustrate the framework’s capabilities and innovations:

• Autonomous Structure Discovery: The model independently learns the most effective data structure configurations, removing the need for predefined, human-designed structures.
• High Precision in Simple and Complex Data Settings: Achieved 99.5% accuracy in ordered ranking for 1D NN search and successfully navigated uniform and high-dimensional data with minimal supervision.
• Efficient Use of Extra Space for Enhanced Accuracy: The framework demonstrated a clear performance boost as additional memory was allocated, showing adaptability in constrained environments.
• Broad Applicability Beyond NN Search: The framework’s flexibility was further highlighted in frequency estimation tasks, where it exceeded the performance of CountMin sketch models in data with Zipfian distributions, indicating potential for other high-demand applications.

In conclusion, this research illustrates a promising step toward the future of machine learning-driven data structure discovery. By harnessing adaptive end-to-end training, this framework efficiently addresses the storage and query challenges that traditional data structures face, especially when working within real-world data constraints. This approach enhances the speed and accuracy of data retrieval and opens avenues for autonomous discovery in data processing, marking a significant advancement in machine learning’s application to structural optimization.

Check out the Paper. All credit for this research goes to the researchers of this project. Also, don’t forget to follow us on Twitter and join our Telegram Channel and LinkedIn Group. If you like our work, you will love our newsletter.. Don’t Forget to join our 55k+ ML SubReddit.

[Sponsorship Opportunity with us] Promote Your Research/Product/Webinar with 1Million+ Monthly Readers and 500k+ Community Members

The post Exploring Adaptive Data Structures: Machine Learning’s Role in Designing Efficient, Scalable Solutions for Complex Data Retrieval Tasks appeared first on MarkTechPost.

The post Exploring Adaptive Data Structures: Machine Learning’s Role in Designing Efficient, Scalable Solutions for Complex Data Retrieval Tasks appeared first on BusinessViewed.

Large language models (LLMs) have revolutionized natural language processing by making strides in text generation, summarization, and translation. Even though they excel at language tasks, they need help handling complex, multi-step reasoning tasks that require careful progression through each step. Researchers have been exploring structured frameworks that enhance these models’ reasoning abilities, moving beyond conventional prompt-based methods.

A significant challenge in advancing LLMs’ capabilities lies in enabling these models to break down and navigate through intricate tasks that involve multiple interconnected steps. Traditional language models often need more attention to critical subtasks within a complex problem, leading to inaccurate or incomplete outcomes. This problem is particularly evident in tasks that demand sequential decision-making or synthesis across various pieces of information. Researchers aim to address this by creating systems that decompose complex tasks into simpler, more manageable parts, enabling models to handle sophisticated tasks more reliably.

Several methods have been proposed to address these challenges, each with its unique approach. Chain-of-thought (CoT) prompting allows models to perform reasoning sequentially, providing prompts that guide step-by-step logic. However, CoT is often limited by its need for manual prompt engineering and needs help with tasks outside its training domain. To enhance this approaches like Tree of Thoughts (ToT) and Graph of Thoughts (GoT) organize problem-solving paths into structured hierarchies, each representing a potential solution route. Despite these advancements, these approaches can become overly intricate for certain problem types, introducing unnecessary complexity into models that perform best with direct prompts.

Researchers from Inria in Paris, France, introduced an innovative framework known as the Tree of Problems (ToP) to overcome these limitations. This method provides a simpler yet effective structure for problem decomposition, focusing on problems that can be divided into analogous subtasks. Unlike the more complex ToT or GoT frameworks, ToP organizes tasks into a hierarchical structure where each node represents a subproblem directly related to the original task. This allows LLMs to solve smaller, similar instances of a larger problem before combining these solutions into a cohesive answer, ultimately reducing computational load and enhancing accuracy.

The ToP framework systematically breaks down a problem into a tree structure composed of simpler sub-tasks. The process starts with a decomposer that divides the primary task into related subtasks and then organizes them within a tree where each node corresponds to a subproblem. A solver, often an LLM configured for task-specific objectives, addresses these atomic problems at the tree’s base. Each node is solved independently, and solutions are merged bottom-up, with the final solution forming at the tree’s root. This method ensures that the LLM only focuses on one problem component at a time, simplifying the reasoning process and minimizing error.

Empirical evaluations have demonstrated the efficiency and performance of the ToP approach, particularly when applied to structured tasks. For example, ToP achieved an accuracy improvement of 40% over GoT in sorting tasks, outperforming CoT and ToT methods by considerable margins. In set intersection tasks, ToP showed an increase in accuracy by 19% over CoT, and in keyword counting tasks, it achieved a 5% improvement, demonstrating its effectiveness across various problem domains. The framework also excelled in tasks such as the Last Letter Concatenation, which recorded higher accuracy rates than CoT in scenarios involving 4, 8, and 16 names. These numbers indicate ToP’s scalability and adaptability across different problem types, making it a promising solution for improving LLM reasoning in complex settings.

Further analysis revealed ToP’s advantages over Least-to-Most (L2M) prompting, another structured approach that involves processing a task step-by-step. In tests with various list lengths, ToP consistently outperformed L2M while requiring fewer computational calls. For lists of 4 and 8 names, ToP achieved comparable or superior accuracy with half as many calls, highlighting its efficiency. On tasks requiring sequential processing, such as coin flipping and object tracking, ToP also demonstrated robustness by handling increased complexity with minimal drop in performance, showing its adaptability for canonical and sequential tasks.

The Tree of Problems framework represents a promising direction for large language model development by addressing key limitations in multi-step reasoning. By breaking down complicated tasks into manageable subproblems and organizing them in a simple, effective tree structure, ToP enhances both accuracy and computational efficiency. This approach outperforms traditional methods and introduces a scalable framework for applying LLMs to more complex reasoning tasks in natural language processing. Through innovations like ToP, LLMs are poised to become more reliable tools in handling diverse, complex tasks, marking a significant step forward in the field.

Check out the Paper. All credit for this research goes to the researchers of this project. Also, don’t forget to follow us on Twitter and join our Telegram Channel and LinkedIn Group. If you like our work, you will love our newsletter.. Don’t Forget to join our 55k+ ML SubReddit.

[Sponsorship Opportunity with us] Promote Your Research/Product/Webinar with 1Million+ Monthly Readers and 500k+ Community Members

The post This AI Paper by Inria Introduces the Tree of Problems: A Simple Yet Effective Framework for Complex Reasoning in Language Models appeared first on MarkTechPost.

The post This AI Paper by Inria Introduces the Tree of Problems: A Simple Yet Effective Framework for Complex Reasoning in Language Models appeared first on BusinessViewed.