Udacity-Bertelsmann New Tech Scholarships

By

-

October 30, 2023

Udacity-Bertelsmann New Tech Scholarships — Image source: Medium

Udacity collaborates with Bertelsmann to offer scholarships for participants to acquire the necessary technology skills. Both organizations have collaborated before for different skills scholarships, but this time, it is for artificial intelligence, cybersecurity, or project management.

The scholarship is called Next Generation Tech Booster and will award thousands of scholarships to students worldwide online. The program offers three specialization courses: Agile Software Developer, Enterprise Security, and AI Programming with Python. Students can choose the course of their choice.

For interested candidates, the application window for this scholarship is already open from October 23, 2023, and will remain open till November 30, 2023. Check out the program details here.

To participate in this program, you need basic knowledge of the field you are choosing and fill in the application form. According to the official Udacity website, “If you’re 18 years or older, have English comprehension, and are looking to develop job-ready skills in project management, artificial intelligence, and cybersecurity, this scholarship program is for you.”

Next Generation Tech Booster is 100% online, and as a learner, you will have access to learning material from a virtual classroom. To graduate, you must complete specified projects within a set of time frames.

This collaboration between Udacity and Bertelsmann to provide scholarships for professionals represents a significant opportunity for the careers of many aspiring tech professionals.

OpenAI Establishes New Team to Combat “Catastrophic Risks” Associated with AI

By

Boudhayan Ghosh

-

October 28, 2023

OpenAI has announced the establishment of a new team dedicated to addressing and minimizing the “catastrophic risks” associated with artificial intelligence.

In their recent update on Thursday, Open AI revealed that this preparedness team’s mission is to diligently monitor, assess, predict, and safeguard against significant challenges arising from AI, which may include threats as those related to nuclear technology.

Additionally, the team will also focus on reducing the potential dangers stemming from “chemical, biological, and radiological threats,” as well as preventing “autonomous replication,” which involves AI independently creating copies of itself. The team will also address risks related to AI’s capacity to deceive humans and tackle issues concerning cybersecurity threats.

Aleksander Madry, currently on leave as the director of MIT’s Centre for Deployable Machine Learning, has been appointed to lead the preparedness team.

OpenAI has underscored that this team will additionally be responsible for crafting and upholding a “risk-informed development policy,” delineating Open AI’s actions for assessing and overseeing AI models.

Will Nightshade Bring a Watershed Moment in the AI-Art Landscape?

By

Boudhayan Ghosh

-

October 27, 2023

Computer Science Professor Ben Zhao at the University of Chicago has developed “Nightshade,” an online tool that can bring a watershed moment in the AI-Art landscape.

The ongoing predicament between artists and AI companies is well known. Whether it be the hollywood writer strike, or the continuing mounting of lawsuits filed against generative AI companies by painters, musicians, etc, for unlawfully scrapping their material from the internet to train AI models is shedding light on the subject’s intricacies.

The question is, how would Nightshade prevent these AI giants from scraping data from the internet? The answer is not in preventing but in poisoning. Nightshade joins the party when an artist uploads their work online but doesn’t want it to be scrapped by AI companies without permission or royalties.

But the question is, what Nightshade specifically do? The AI tool alters the information related to a particular image, eventually making the image-generating AI models learn the wrong names of things they are looking at. This concept is called “data poisoning.”

What is Nightshade? It's a tool that performs a data poisoning attack against generative AI image models. Poisoning is not new. Poisoning genAI models at scale is new. You can read the MIT TR article for the high level story. For details, here's the paper: https://t.co/0mIZgOl1Fp
— Glaze at UChicago (@TheGlazeProject) October 24, 2023

Ben Zhao and his team poisoned images of dogs by embedding information in the pixels to deceive an AI model into perceiving them as cats. In their experiment, the researchers put this attack to the test on Stable Diffusion’s latest models. By positioning just 50 images of dogs and then instructing Stable Diffusion to generate dog images, the results turned unusual, producing creatures with excessive limbs and cartoonish features. When the number of poisoned samples increased to 300, an attacker could successfully coerce Stable Diffusion into generating dog images that resembled cats.

ok ok, 1 last tweet I promise.
I realized the most surprising result was not included in the MIT TR article. You can read the details in the paper (fig17), and I will just leave the figure here. FIN/ pic.twitter.com/zeDDlHbVEO
— Glaze at UChicago (@TheGlazeProject) October 24, 2023

Furthermore, Nightshade’s method of poisoning data poses a formidable challenge in terms of defense. This is because it necessitates that developers of AI models meticulously identify and remove images containing altered pixels. These pixels are intentionally designed to be inconspicuous to the human eye and may prove challenging even for data-scrapping software tools to detect.

The team plans to release Nightshade as an open-source tool, enabling others to experiment and create their own variations. Zhao explains that the tool’s potency increases as more people utilize it as large AI models rely on datasets that can encompass billions of images; the more poisoned images that find their way into the model, the more significant the potential impact of this technique.

So next time, all an artist needs to do before uploading an image online is to use Nightshade and poison the material so that AI models like DALLE and Midjourney can’t identify the true nature of the picture, eventually making the data deceptive and problematic for the system.

China Accepts Invitation to the First Ever Global AI Summit Hosted by UK

By

Boudhayan Ghosh

-

October 27, 2023

China UK Global AI Summit — Source: THE CEO PUBLICATION

British Deputy Prime Minister Oliver Dowden has confirmed that China will take part in the UK’s global AI summit scheduled for next week.

Chinese officials have only received invitations for two days of the summit, which will be held on November 1-2, and not the second-day meeting, which is focused on AI safety and security risks.

The invitation to China has triggered a strong negative response, with figures like former British Prime Minister Liz Truss advocating for China’s exclusion. Concerts have been raised regarding Beijing’s utilization of AI for state control and national security.

In a letter addressed to the current Prime Minister, Rishi Sunak, Truss stated, “No reasonable person expects China to abide by anything agreed to at this kind of summit.”

However, Sunak, who extended the invitation to China, has defended this decision. He asserts that involving all major AI powers is indispensable for developing a robust AI strategy.

Sunak’s decision to include China in the UK global AI summit was embraced by Arati Prabhakar, U.S. President Joe Biden’s top science adviser. She underscored the importance of fostering a global conversation on the matter.

SurveyMonkey Launched ‘Build with AI’ to Create Instant Surveys

By

Boudhayan Ghosh

-

October 26, 2023

SurveyMonkey Build with AI survey — Source: SurveyMonkey

SurveyMonkey has launched a feature named “Build with AI,” which employs AI technology to auto-generate surveys once users outline their objectives and target audience.

The salient objective of Build with AI is to significantly reduce the time needed to create tailored polls and questionnaires.

Using this feature, SurveyMonkey users can craft surveys with the help of generative AI using Open AI’s advanced GPT-3.5-turbo models.

SurveyMoneky’s official account thoroughly describes how to build a survey. First, a user must click “Create Survey” and select “Build with AI.” Next, input a prompt and then click on the paper airplane icon.

Following this, a user will be prompted to “Create Survey” once again. At this stage, one can take a moment to review the survey. If satisfied with the result, click on “Use Survey.” This will allow users to customize the survey according to their preferences.

Additional features include assistance in survey creation, ensuring that users can pose pertinent questions effectively regardless of expertise. It also incorporates safety measures to safeguard data when employing GPT-3, offers support for over 50 languages, and seamlessly integrates with SurveyMonkey’s existing AI features, facilitating comprehensive data collection and analysis through AI.

Motorola’s Slap Bracelet Smartphone is the New Design Standard

By

Boudhayan Ghosh

-

October 26, 2023

Motorola slap bracelet smartphone — Source: Motorola

Motorola’s new product is straight from the future. Keeping up with the technological innovation madness of the 21st century, the Lenovo-owned company showcased its latest smartphone that can wrap around your wrist like a slap bracelet.

Motorola released its avant-garde concept during the Lenovo Tech World forum, showcasing the smartphone’s non-rigid facets. The innovative conceptual device employs an FHD + pOLED display that can be flexed and molded into various shapes based on users’ requirements.

For instance, it can serve as a regular Android phone, stand independently, or transform into a wrist-worn watch accessory.

Besides the elastic smartphone, Motorola also unveiled the latest AI features they have incorporated into their smartphones. One of them is a generative AI model running locally on the device, enabling users to personalize their smartphones with AI-generated wallpapers that match their outfit selection.

MotoAI, a personnel voice and text assistant that harnesses the capabilities of a Large Language Model (LLM) and runs on smartphones and PCs, was also released during the forum. This feature can manage schedules, respond to queries, and compose messages. Motorola also unveiled the mobile Doc Scanner, which gives crisp and clear images by minimizing wrinkles and shadows.

Finally, features like AI Text Summarization, which segregates necessary information from long-form chats, emails, or reports and concise them for better understanding, and Privacy Content Obfuscation, which blurs profile pictures and names in social posts to safeguard users information and privacy were also released.

Amazon’s New Generative AI Tool for Advertisers

By

Boudhayan Ghosh

-

October 26, 2023

Amazon Generative AI advertisers — Source: Amazon

Amazon Ads has launched a new generative AI tool in beta to help advertisers improve their ads performance.

According to Amazon executives, creating captivating and authentic advertising content might be costly, often needing additional expertise, something that small-scale advertisers need help with. Keeping this in mind, Amazon wants to help advertisers with their ad campaigns by incorporating the new image generation tool within their system.

Colleen Aubrey, senior vice president of Amazon Ads products and technology, says their new image generator is simple and easy to use and, at the same time, gives better outcomes. He further states that not only would advertisers benefit from it, but customers would also see more “engaging and visually rich” ads.

All an advertiser needs to do is go to the Amazon Ad Console, choose their product, and click “Generate.” Using generative AI, this tool promptly produces a collection of lifestyle and brand-focused images tailored to the product specifications.

These images can be further enhanced by inputting concise text prompts, allowing for the rapid creation and testing of multiple versions to fine-tune performance optimization.

Amazon also claims that products showcased in a lifestyle context can achieve a 40% increase in click-through rates compared to advertisements featuring typical product images.

Carry Your AI Everywhere With Humane’s AI Pin

By

Boudhayan Ghosh

-

October 25, 2023

Humane has announced that their first intelligent clothing-based wearable device, the “Humane AI Pin,” is launching on November 9^th.

Coming straight from the horse’s mouth, the Humane AI Pin is engineered to integrate seamlessly with one’s unique style and clothing choices. The wearable device also employs a variety of sensors to facilitate natural and instinctive computing.

The AI pin, which operates independently without requiring connection to a smartphone or any other companion device, features AI-driven optical recognition and a laser-protected display, both powered by the advanced Snapdragon platform from Qualcomm technologies.

According to Humane, privacy is of utmost importance to the device. Features like the absence of a wake word eliminating an “always on listening” align with Humane’s commitment to creating products prioritizing trust and security. Humane AI Pin also has a “trust light,” which lights up whenever the wearable device records data.

The device, which made its way to the Time Magazines “Best Inventions of 2023,” also debuted at the Paris Fashion Week as Humane AI Pin collaborated with Paris-based fashion house Coperni.

Our Ai Pin made the “@TIME Best Inventions” list. Thank you to our team, partners, and investors across the planet. See you on November 9th! https://t.co/S7xZsDFD32
— Humane (@Humane) October 24, 2023

Meta Announces Tools to Advance Socially Intelligent Robots

By

Jalaj Jain

-

October 24, 2023

Meta Announces AI tools to advance socially intelligent robots — Image source: meta

Facebook AI Research (FAIR) announced the launch of Habitat 3.0, Habitat Synthetic Scenes Dataset (HSSD-200), and HomeRobot, giving AI-powered assistants a whole new level of meaning. Meta aims to develop socially intelligent robots that adapt to human preferences and assist with everyday tasks, underscoring the critical role of embedded systems in shaping the future of AR and VR experiences.

These tools are launched to overcome all the challenges that usual AI-powered assistants face, including scalability, standardized benchmarking, collaborative robotics, and safety concerns.

FAIR’s Habitat 3.0 enables large-scale training for human-robot interaction in realistic indoor environments. AI agents trained using these simulations can demonstrate collaborative behaviors, from navigating narrow corridors to efficiently dividing tasks.

And, the introduction of the HSSD-200 addresses critical issues in training robots in simulated environments. It offers highly detailed 3D environments, fine-grained semantic categorization, and efficient asset compression. FAIR’s experiments have shown that HSSD-200’s smaller yet higher-quality dataset can lead to AI agents with comparable or superior performance in real-world scenes.

To help things up in collaborative robotics, FAIR built HomeRobots, a software stack for physical and simulated autonomous manipulation. It consists of benchmarks, baseline behaviors, and interfaces that focus on tasks like delivering requested objects.

Habitat 3.0, HSSD-200, and HomeRobots represent a significant step toward developing socially embodied AI agents to collaborate with and assist humans in their everyday lives. The work of FAIR can potentially bridge the gap between simulation and the physical world, opening new frontiers in human-robot collaboration and interaction.

NVIDIA Introduces New AI Agent to Revolutionize Robot Learning

By

Jalaj Jain

-

October 23, 2023

NVIDIA Introduces New AI Agent to Revolutionize Robot Learning — Image Source: multiplatform.ai

NVIDIA Research has made a new AI tool called Eureka that teaches robots to do complex tasks. It taught a robot how to do quick pen spin like a human for the first time.

According to the official blog post of NVIDIA, it is just one of the 30 tasks the robots have learned to perform expertly. Besides pen spinning, it has taught many tasks to robots, such as opening drawers, manipulating scissors, and tossing and catching balls.

Eureka teaches robots by writing reward algorithms to train bots autonomously. The tool leverages the cutting-edge capabilities of GPT-4’s large language model (LLM) and NVIDIA’s own Isaac Gym, a development platform for building 3D tools and applications.

Importantly, this tool works automatically and does not need human prompting or pre-defined templates. It teaches robots by formulating a reward algorithm specifically tailored to the task and morphology of the robot. The reward scheme allows robots to learn by trial and error.

When tested against human specialists in 29 different environments using ten different robot platforms, Eureka produced superior reward programs. The official paper of NVIDIA mentions this tool leads to an average performance improvement of more than 50% for the bots.

The code and benchmarks of Eureka will be open-source. This will allow developers to use resources to expand on the research and leverage the community. Recently, it released a video that shows the workings of this tool.

The introduction of Eureka by NVIDIA has the potential to revolutionize robot learning by providing a platform for robots to learn complex skills with high efficiency. As the technology develops more, we expect an increasing integration of these AI-driven robots into various industries.