Daily Tech Digest - February 19, 2024

Why artificial general intelligence lies beyond deep learning

Decision-making under deep uncertainty (DMDU) methods such as Robust Decision-Making may provide a conceptual framework to realize AGI reasoning over choices. DMDU methods analyze the vulnerability of potential alternative decisions across various future scenarios without requiring constant retraining on new data. They evaluate decisions by pinpointing critical factors common among those actions that fail to meet predetermined outcome criteria. The goal is to identify decisions that demonstrate robustness — the ability to perform well across diverse futures. While many deep learning approaches prioritize optimized solutions that may fail when faced with unforeseen challenges (such as optimized just-in-time supply systems did in the face of COVID-19), DMDU methods prize robust alternatives that may trade optimality for the ability to achieve acceptable outcomes across many environments. DMDU methods offer a valuable conceptual framework for developing AI that can navigate real-world uncertainties. Developing a fully autonomous vehicle (AV) could demonstrate the application of the proposed methodology. The challenge lies in navigating diverse and unpredictable real-world conditions, thus emulating human decision-making skills while driving. 


Bouncing back from a cyber attack

In the case of a cyber attack, the inconceivable has already happened – all you can do now is bounce back. The big picture issue is that too often IoT (internet of things) networks are filled with bad code, poor data practices, lack of governance, and underinvestment in secure digital infrastructure. Due to the popularity and growth of IoT, manufacturers of IoT devices spring up overnight promoting products that are often constructed using lower-quality components and firmware, which can have sometimes well-known vulnerabilities exposed due to poor design and production practices. These vulnerabilities are then introduced to a customer environment increasing risk and possibly remaining unidentified. So, there’s a lot of work to do, including creating visibility over deep, widely connected networks with a plethora of devices talking to each other. All too often, IT and OT networks run on the same flat network. For these organisations, many are planning segmentation projects, but they are complex and disruptive to implement, so in the meantime companies want to understand what's going on in these environments and minimise disruption in the event of an attack.


Diversity, Equity, and Inclusion for Continuity and Resilience

As continuity professionals, the average age tends to skew older, so how do we continue to bring new people to the fold to ensure they feel like they can learn and be respected in the industry? Students need to be made aware this is an industry they can step into. Unfortunately, many already have experience seeing active shooter drills as the norm. They may have never organized one, but they have participated in many of these drills in school. Why not take advantage of that experience for the students who are interested in this field? Taking their advice could make exercising like active shooter or weather events less traumatic. Listening to their experience – doing it for at least 13 years – gives them a lot of insight from even Millennials who grew up at the forefront of school shootings, but not actively exercising what to do if it happens while in school. These future colleagues’ insights could change how we do specific exercises and events to benefit everyone. Still, there must be openness to new and fresh ideas and treating them with validity instead pushing them off due to their age and experience. Similarly, people with disabilities have always been vocal about their needs. 


AI’s pivotal role in shaping the future of finance in 2024 and beyond

As AI becomes more embedded in the financial fabric, regulators are crafting a nuanced framework to ensure ethical AI use. The Reserve Bank of India (RBI) and the Securities and Exchange Board of India (SEBI) have initiated guidelines for responsible AI adoption, emphasising transparency, accountability, and fairness in algorithmic decision-making processes. While the benefits are palpable, challenges persist. The rapid pace of AI integration demands a strategic approach to ensure a safe, financial eco-system ... The evolving nature of jobs due to AI necessitates a concerted effort towards upskilling the workforce. A McKinsey Global Institute report indicates that approximately 46% of India’s workforce may undergo significant changes in their job profiles due to automation and AI. To address this, collaborative initiatives between the government, educational institutions, and the private sector are imperative to equip the workforce with the requisite skills for the future. ... The Reserve Bank of India (RBI) and the Securities and Exchange Board of India (SEBI) have recognised the need for ethical AI use in the financial sector. Establishing clear guidelines and frameworks for responsible AI governance is crucial. 


How to proactively prevent password-spray attacks on legacy email accounts

Often with an ISP it’s hard to determine the exact location from which a user is logging in. If they access from a cellphone, often that geographic IP address is in a major city many miles away from your location. In that case, you may wish to set up additional infrastructure to relay their access through a tunnel that is better protected and able to be examined. Don’t assume the bad guys will use a malicious IP address to announce they have arrived at your door. According to Microsoft, “Midnight Blizzard leveraged their initial access to identify and compromise a legacy test OAuth application that had elevated access to the Microsoft corporate environment. The actor created additional malicious OAuth applications.” The attackers then created a new user account to grant consent in the Microsoft corporate environment to the actor-controlled malicious OAuth applications. “The threat actor then used the legacy test OAuth application to grant them the Office 365 Exchange Online full_access_as_app role, which allows access to mailboxes.” This is where my concern pivots from Microsoft’s inability to proactively protect its processes to the larger issue of our collective vulnerability in cloud implementations. 


How To Implement The Pipeline Design Pattern in C#

The pipeline design pattern in C# is a valuable tool for software engineers looking to optimize data processing. By breaking down a complex process into multiple stages, and then executing those stages in parallel, engineers can dramatically reduce the processing time required. This design pattern also simplifies complex operations and enables engineers to build scalable data processing pipelines. ...The pipeline design pattern is commonly used in software engineering for efficient data processing. This design pattern utilizes a series of stages to process data, with each stage passing its output to the next stage as input. The pipeline structure is made up of three components: The source: Where the data enters the pipeline; The stages: Each stage is responsible for processing the data in a particular way; The sink: Where the final output goes Implementing the pipeline design pattern offers several benefits, with one of the most significant benefits in efficiency of processing large amounts of data. By breaking down the data processing into smaller stages, the pipeline can handle larger datasets. The pattern also allows for easy scalability, making it easy to add additional stages as needed. 


Accuracy Improves When Large Language Models Collaborate

Not surprisingly, this idea of group-based collaboration also makes sense with large language models (LLMs), as recent research from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) is now showing. In particular, the study focused on getting a group of these powerful AI systems to work with each other using a kind of “discuss and debate” approach, in order to arrive at the best and most factually accurate answer. Powerful large language model AI systems, like OpenAI’s GPT-4 and Meta’s open source LLaMA 2, have been attracting a lot of attention lately with their ability to generate convincing human-like textual responses about history, politics and mathematical problems, as well as producing passable code, marketing copy and poetry. However, the tendency of these AI tools to “hallucinate”, or come up with plausible but false answers, is well-documented; thus making LLMs potentially unreliable as a source of verified information. To tackle this problem, the MIT team claims that the tendency of LLMs to generate inaccurate information will be significantly reduced with their collaborative approach, especially when combined with other methods like better prompt design, verification and scratchpads for breaking down a larger computational task into smaller, intermediate steps.


There's AI, and Then There's AGI: What You Need to Know to Tell the Difference

For starters, the ability to perform multiple tasks, as an AGI would, does not imply consciousness or self-will. And even if an AI had self-determination, the number of steps required to decide to wipe out humanity and then make progress toward that goal is too many to be realistically possible. "There's a lot of things that I would say are not hard evidence or proof, but are working against that narrative [of robots killing us all someday]," Riedl said. He also pointed to the issue of planning, which he defined as "thinking ahead into your own future to decide what to do to solve a problem that you've never solved before." LLMs are trained on historical data and are very good at using old information like itineraries to address new problems, like how to plan a vacation. But other problems require thinking about the future. "How does an AI system think ahead and plan how to eliminate its adversaries when there is no historical information about that ever happening?" Riedl asked. "You would require … planning and look ahead and hypotheticals that don't exist yet … there's this big black hole of capabilities that humans can do that AI is just really, really bad at."


Metaverse and the future of product interaction

As the metaverse continues to evolve, so must the approach to product design. This includes considering how familiar objects can be repurposed as functional interface elements in a virtual environment. Additionally, understanding the dynamics of group interactions in virtual spaces is crucial. Designers must anticipate these trends and adapt their designs accordingly, ensuring that products remain relevant and engaging in the ever-changing landscape of the metaverse. In India, the metaverse presents significant opportunities for businesses to redefine consumer experiences. It opens up possibilities for more interactive, personalised, and adventurous engagements with customers. This not only increases customer engagement and loyalty but also creates new avenues for value exchange and revenue streams. The metaverse, with its potential to impact diverse sectors like communications, retail, manufacturing, education, and banking, is poised to be a game-changer in the Indian market. ... As the metaverse continues to expand its reach and influence, businesses and designers in India and around the world must evolve to meet the demands of this new digital era.


Build trust to win out with genAI

Businesses need to adopt ‘responsible technology’ practices, which will give them a powerful lever that enables them to deploy innovative genAI solutions while building trust with consumers. Responsible tech is a philosophy that aligns an organization’s use of technology to both individuals’ and society’s interests. It includes developing tools, methodologies, and frameworks that observe these principles at every stage of the product development cycle. This ensures that ethical concerns are baked in at the outset. This approach is gaining momentum, as people realize how technologies such as genAI, can impact their daily lives. Even organizations such as the United Nations are codifying their approach to responsible tech. Consumers urgently want organizations to be responsible and transparent with their use of genAI. This can be a challenge because, when it comes to transparency, there are a multitude of factors to consider, including everything from acknowledging AI is being used to disclosing what data sources are used, what the steps were taken to reduce bias, how accurate the system is, or even the carbon footprint associated with the genAI system.



Quote for the day:

"Entrepreneurs average 3.8 failures before final success. What sets the successful ones apart is their amazing persistence." -- Lisa M. Amos

No comments:

Post a Comment