The African hawk-eagle (Aquila spilogaster) is a large bird of prey. Like all eagles, it belongs to the family Accipitridae. The species's feathered legs mark it as a member of the subfamily Aquilinae. The African hawk-eagle breeds in tropical sub-Saharan Africa. It is a bird of assorted woodland, including both savanna and hilly areas, but they tend to occur in typically dry woodland. The species tends to be rare in areas where their preferred habitat type is absent. The African hawk-eagle is powerfully built and hunts small to medium-sized mammals and birds predominantly, occasionally taking reptiles and other prey as well. This African hawk-eagle perching on a branch was photographed in Damaraland, Namibia. Photograph credit: Charles J. Sharp Recently featured: Christ Crowned with Thorns Rambutan Galaxea fascicularis Archive More featured pictures

April 19: Primrose Day in London Aftermath of the Oklahoma City bombing 1713 – With no living male heirs, Charles VI, Holy Roman Emperor, issued the Pragmatic Sanction, allowing the Habsburg hereditary possessions to be inherited by a daughter. 1809 – War of the Fifth Coalition: French general Louis-Nicolas Davout defeated an Austrian force in Lower Bavaria, allowing him to rejoin the main French army. 1987 – The fictional Simpson family made their first appearance in the short "Good Night", aired in a segment of the The Tracey Ullman Show. 1995 – A truck bombing destroyed much of the Alfred P. Murrah Federal Building (aftermath pictured) in Oklahoma City, killing 168 people and injuring more than 680 others. 2015 – Freddie Gray, a 25-year-old African American, died of injuries sustained while in the custody of the Baltimore Police Department. Uesugi Kenshin (d. 1578)Elizabeth Dilling (b. 1894)Denis O'Brien (b. 1958) More anniversaries: April 18 April 19 April 20 Archive By email List of days of the year About

Rembrandt’s Stunning Sketch of a Lion Will Be Sold at Auction to Fund Wildlife Conservation Efforts “Young Lion Resting” is among dozens of Dutch Golden Age artworks from the Leiden Collection that are now on view at the H’ART Museum in Amsterdam Young Lion Resting, Rembrandt van Rijn, circa 1638-42 The Leiden Collection A new exhibition in Amsterdam brings together 75 paintings from the Leiden Collection, a stunning assortment of art from the Dutch Golden Age. Highlights of the show include 18 artworks by Rembrandt van Rijn, including a drawing of a lion that will be sold to fund wildlife conservation efforts. The H’ART Museum’s “From Rembrandt to Vermeer, Masterpieces From the Leiden Collection” marks the first time that so many works from the Leiden Collection—including all 18 of its Rembrandts—have been publicly displayed in the Netherlands. Owned by collectors Thomas and Daphne Kaplan, the more than 200 paintings and drawings represent one of the world’s largest private collections of 17th-century Dutch art, according to a statement. “The idea was born to bring all the Rembrandts from this remarkable collection to the Dutch capital in celebration of Amsterdam’s 750th anniversary,” Annabelle Birnie, the H’ART Museum’s general director, tells DutchNews’ Senay Boztas. “It is thanks to this special partnership [with the Leiden Collection] that we are able to share these masterpieces with the city and the Netherlands.” Study of a Woman in a White Cap, Rembrandt van Rijn, circa 1640 The Leiden Collection Born in the Dutch city of Leiden in 1606, Rembrandt developed a reputation as one of the Netherlands’ most famous artists during his lifetime. He’s known for his dramatic treatment of light and shadows, particularly in group portraits like The Anatomy Lesson of Dr. Nicolaes Tulp (1632) and The Night Watch (1642). “The character and spirit of the Dutch people, as seen through portraiture, constitutes a major component of this exhibition,” says the museum in the statement. “One continually senses the indomitable spirit of the Dutch people as one walks through the show, not only when marveling at Rembrandt’s formal portraits of wealthy burgers, but also when viewing his expressive tronies (character studies) of men and women from a range of social classes.” Rembrandt is also famous for his extensive collection of self-portraits, which were coveted by many collectors of his day. As Thomas Kaplan tells DutchNews, the Dutch master “painted probably more self-portraits than any other artist.” Self-Portrait with Shaded Eyes, Rembrandt van Rijn, 1634 The Leiden Collection “It’s not because he was an egomaniac. It’s because he was a celebrity, and his portraits sold well,” Kaplan adds. “If you had a self-portrait of Rembrandt on your wall, it was as iconic to someone from the Netherlands as having an Andy Warhol Campbell’s Soup Can, or a Cy Twombly.” However, Rembrandt didn’t only draw portraits—and sometimes he even drew animals. Young Lion Resting (1638-42), which is now on display at the H’ART, depicts a big cat lying down, wearing a leash around its neck. Its front paws are crossed, and its wide eyes face forward. Young Lion Resting was the first Rembrandt work Kaplan purchased. Its subject is especially significant to the collector, who co-founded Panthera, an organization dedicated to preserving seven species of big cats, in 2006. Next year, he plans to sell Young Lion Resting at auction and donate the proceeds to the organization. Young Woman Seated at a Virginial, Johannes Vermeer, circa 1670-75 The Leiden Collection “Wildlife conservation is the one passion I have which surpasses Rembrandt—and I want to attract more people to that cause,” he tells the Art Newspaper’s Martin Bailey. Rembrandt likely observed the lion at an Amsterdam menagerie—a traveling exhibition of animals brought from North Africa to the Netherlands on trading ships of the Dutch East India Company. Per the Art Newspaper, the artist may have later used the animal’s likeness in portraits of St. Jerome, who is often depicted alongside a lion (as he supposedly removed a thorn from its paw). As Kaplan tells the Art Newspaper, Rembrandt “gives a greater interior life to a cat than most artists can to a human.” “From Rembrandt to Vermeer, Masterpieces From the Leiden Collection” will be on view at the H’ART Museum in Amsterdam through August 24, 2025. Get the latest stories in your inbox every weekday.

Google has surged ahead in the enterprise AI race after perceived stumbles. VentureBeat details the Gemini models, TPU advantage & agent ecosystem driving its turnaround.Read More

In a letter to the Social Security Administration’s Inspector General’s office requesting an investigation into DOGE, Ranking Member Rep. Gerald Connolly (D-VA) alleged that the government entity created by Elon Musk supposedly to reduce the size of the federal government is now constructing a “cross-agency master database” of sensitive personal information.Wired appeared to back up Connolly’s allegations on Friday, detailing an effort at DOGE to fold this database into the Department of Homeland Security, the counterterrorism agency founded after 9/11. Specifically, “mass amounts” of personal data harvested from the IRS, SSA, and voting records in Pennsylvania and Florida were recently uploaded into servers at the United States Customs and Immigration Services (USCIS), which processes immigration cases.Connolly cited testimony from SSA whistleblowers who witnessed DOGE engineers accessing the agency’s IT system with “backpacks full of laptops, each with access to different agency systems”, with the aim of combining them into one database. Connolly warned that not only would such a database pose a threat to government cybersecurity, which siloes its information across several agencies to prevent cyberattacks from accessing all information at once, it was also very likely violating several privacy laws.The Committee has also received reports about troubling, fumbling efforts by DOGE to combine sensitive information held by SSA, the IRS, HHS, and other agencies into a single cross-agency master database. Improving how federal agencies share data to improve outcomes and customer service is a longstanding and bipartisan goal in Congress. Information obtained by the Committee, however, indicates that DOGE is carrying out its work in a manner that disregards important cybersecurity and privacy considerations, potentially in violation of the law. In an apparent attempt to sidestep network security controls, the Committee has learned that DOGE engineers have tried to create specialized computers for themselves that simultaneously give full access to networks and databases across different agencies. Such a system would pose unprecedented operational security risks and undermine the zero-trust cybersecurity architecture that prevents a breach at one agency from spreading across the government. Information obtained by the Committee also indicates that individuals associated with DOGE have assembled backpacks full of laptops, each with access to different agency systems, that DOGE staff is using to combine databases that are currently maintained separately by multiple federal agencies.Though several other House investigations into DOGE’s activities have revealed their data harvesting efforts at other agencies, Rep. Connolly’s letter is the first to allege that DOGE, the government-shrinking agency founded by Elon Musk and now wreaking havoc across the federal government, is pooling everyone’s data into one giant database. “I have long championed efforts to improve data sharing across the government to combat improper payments and to increase government efficiency,” Connolly wrote. “But any efforts to reform our current systems must be undertaken with the utmost sensitivity and concern for privacy, security, and the Social Security payments that millions of people rely on.”The privacy implications of this cross-agency database would not only violate numerous privacy laws and cause cybersecurity risks, but could also become a potent weapon for whomever can access said database: it could be used by the government to conduct mass surveillance of whomever they wished to target, such as immigrants, or become a target for outside actors seeking a trove of private and personal data. Either way, “it’s terrifying,” John Davisson, senior counsel and director of litigation at the Electronic Privacy Information Center, told Nextgov, adding that this was exactly what the Privacy Act was supposed to prevent. “Aggregation of data is building a weapon, essentially, and it’s one that can be used in a lot of different ways.” See More:

Despite rapid advances in vision-language modeling, much of the progress in this field has been shaped by models trained on proprietary datasets, often relying on distillation from closed-source systems. This reliance creates barriers to scientific transparency and reproducibility, particularly for tasks involving fine-grained image and video understanding. Benchmark performance may reflect the training data and black-box model capabilities more than architectural or methodological improvements, making it difficult to assess true research progress. To address these limitations, Meta AI has introduced the Perception Language Model (PLM), a fully open and reproducible framework for vision-language modeling. PLM is designed to support both image and video inputs and is trained without the use of proprietary model outputs. Instead, it draws from large-scale synthetic data and newly collected human-labeled datasets, enabling a detailed evaluation of model behavior and training dynamics under transparent conditions. The PLM framework integrates a vision encoder (Perception Encoder) with LLaMA 3 language decoders of varying sizes—1B, 3B, and 8B parameters. It employs a multi-stage training pipeline: initial warm-up with low-resolution synthetic images, large-scale midtraining on diverse synthetic datasets, and supervised fine-tuning using high-resolution data with precise annotations. This pipeline emphasizes training stability and scalability while maintaining control over data provenance and content. A key contribution of the work is the release of two large-scale, high-quality video datasets addressing existing gaps in temporal and spatial understanding. The PLM–FGQA dataset comprises 2.4 million question-answer pairs capturing fine-grained details of human actions—such as object manipulation, movement direction, and spatial relations—across diverse video domains. Complementing this is PLM–STC, a dataset of 476,000 spatio-temporal captions linked to segmentation masks that track subjects across time, allowing models to reason about “what,” “where,” and “when” in complex video scenes. Technically, PLM employs a modular architecture that supports high-resolution image tiling (up to 36 tiles) and multi-frame video input (up to 32 frames). A 2-layer MLP projector connects the visual encoder to the LLM, and both synthetic and human-labeled data are structured to support a wide range of tasks including captioning, visual question answering, and dense region-based reasoning. The synthetic data engine, built entirely using open-source models, generates ~64.7 million samples across natural images, charts, documents, and videos—ensuring diversity while avoiding reliance on proprietary sources. Meta AI also introduces PLM–VideoBench, a new benchmark designed to evaluate aspects of video understanding not captured by existing benchmarks. It includes tasks such as fine-grained activity recognition (FGQA), smart-glasses video QA (SGQA), region-based dense captioning (RDCap), and spatio-temporal localization (RTLoc). These tasks require models to engage in temporally grounded and spatially explicit reasoning. Empirical evaluations show that PLM models, particularly at the 8B parameter scale, perform competitively across 40+ image and video benchmarks. In video captioning, PLM achieves gains of +39.8 CIDEr on average over open baselines. On PLM–VideoBench, the 8B variant closes the gap with human performance in structured tasks such as FGQA and shows improved results in spatio-temporal localization and dense captioning. Notably, all results are obtained without distillation from closed models, underscoring the feasibility of open, transparent VLM development. In summary, PLM offers a methodologically rigorous and fully open framework for training and evaluating vision-language models. Its release includes not just models and code, but also the largest curated dataset for fine-grained video understanding and a benchmark suite that targets previously underexplored capabilities. PLM is positioned to serve as a foundation for reproducible research in multimodal AI and a resource for future work on detailed visual reasoning in open settings. Here is the Paper, Model and Code. Also, don’t forget to follow us on Twitter and join our Telegram Channel and LinkedIn Group. Don’t Forget to join our 90k+ ML SubReddit. Asif RazzaqWebsite |  + postsBioAsif Razzaq is the CEO of Marktechpost Media Inc.. As a visionary entrepreneur and engineer, Asif is committed to harnessing the potential of Artificial Intelligence for social good. His most recent endeavor is the launch of an Artificial Intelligence Media Platform, Marktechpost, which stands out for its in-depth coverage of machine learning and deep learning news that is both technically sound and easily understandable by a wide audience. The platform boasts of over 2 million monthly views, illustrating its popularity among audiences.Asif Razzaqhttps://www.marktechpost.com/author/6flvq/An In-Depth Guide to Firecrawl Playground: Exploring Scrape, Crawl, Map, and Extract Features for Smarter Web Data ExtractionAsif Razzaqhttps://www.marktechpost.com/author/6flvq/Meta AI Introduces Perception Encoder: A Large-Scale Vision Encoder that Excels Across Several Vision Tasks for Images and VideoAsif Razzaqhttps://www.marktechpost.com/author/6flvq/IBM Releases Granite 3.3 8B: A New Speech-to-Text (STT) Model that Excels in Automatic Speech Recognition (ASR) and Automatic Speech Translation (AST)Asif Razzaqhttps://www.marktechpost.com/author/6flvq/A Hands-On Tutorial: Build a Modular LLM Evaluation Pipeline with Google Generative AI and LangChain

11 Docker Container Images for Generative AI & ML Projects 0 like April 18, 2025 Share this post Author(s): Youssef Hosni Originally published on Towards AI. Docker containers offer significant advantages for machine learning by ensuring consistent, portable, and reproducible environments across different systems. By encapsulating all dependencies, libraries, and configurations in a container, Docker eliminates compatibility issues and the “it works on my machine” problem. This makes it easier to move ML projects between development, cloud, or production environments without worrying about differences in setup. Additionally, Docker enables scalability and isolation, allowing machine learning workflows to be easily scaled using tools like Kubernetes, and ensuring that dependencies do not conflict between different projects. In this article, we will explore 11 Docker container images for Generative AI and machine learning projects. These include tools for development environments, deep learning frameworks, machine learning lifecycle management, workflow orchestration, and large language models. I. Machine Learning & Data Science PythonJupyter Notebook data science stack II. Generative AI & Deep Learning 3. Hugging Face Transformers 4. NVIDIA CUDA deep learning runtime 5. TensorFlow 6. PyTorch 7. Ollama 8. Qdrant III. Workflow Orchestration & ML Lifecycle Management 9. Airflow 10. MLflow 11. Kubeflow Notebooks Most insights I share in Medium have previously been shared in my weekly newsletter, To Data & Beyond. If you want to be up-to-date with the frenetic world of AI while also feeling inspired to take action or, at the very least, to be well-prepared… Read the full blog for free on Medium. Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor. Published via Towards AI Towards AI - Medium Share this post

Avatar: The Last Airbender is one of the greatest cartoons of all time. The story of Aang and friends has been adapted into a movie, a live-action series, and multiple video games. But the end of that story isn't actually the end of the story. There have since been a series of graphic novels that continue Aang's journey after the fall of Firelord Ozai and before the start of The Legend of Korra.There have been five Avatar: The Last Airbender graphic novels released since the end of the animated series, and a new box set was just released that features all of them. Not only that, Amazon just recently dropped the price down to $83 in a historic new low.Avatar: The Last Airbender Omnibus Boxed Set SaleNew ReleaseAvatar: The Last Airbender Omnibus Boxed SetThe ATLA graphic novels were written by award winning cartoonist Gene Luen Yang and feature artwork from Gurihiru. The stories were created in collaboration with the original creators of the beloved Nickelodeon series and act as a bridge between the end of The Last Airbender and the start of The Legend of Korra. The newly released Omnibus features the complete stories of the following:Avatar: The Last Airbender - The PromiseAvatar: The Last Airbender - The SearchAvatar: The Last Airbender - The RiftAvatar: The Last Airbender - Smoke and ShadowAvatar: The Last Airbender - North and SouthWith the current sale on Amazon, this new release is the most affordable way to purchase the entire series. This set had already gone on sale several times during its preorder period, but this new discount is the first time it's dropped this far below its normal retail price. If you love the ATLA series and haven't already read the continuation of the story, this price cut makes it a great time to do so.See more Avatar: The Last Airbender stuff:Avatar: The Last Airbender - The Complete SeriesSee it at Amazon16” Appa PlushSee it at AmazonAvatar the Last Airbender: 18 Book Box SetSee it at AmazonToph Funko PopSee it at AmazonAre There Any Other Upcoming ATLA Stories?Fans will be happy to know that new episodes from the Avatar universe are definitely on their way. Avatar: The Last Airbender creators Michael DiMartino and Bryan Konietzko recently announced a new series called Avatar: Seven Havens, which will be a 26-episode, 2D animated series following a young Earthbender who is the next Avatar following Korra.In a press release, Nickelodeon says Seven Havens will be set in a world “shattered by a devastating cataclysm. A young Earthbender discovers she’s the new Avatar after Korra - but in this dangerous era, that title marks her as humanity’s destroyer, not its savior. Hunted by both human and spirit enemies, she and her long-lost twin must uncover their mysterious origins and save the Seven Havens before civilization’s last strongholds collapse.”Avatar Studios is also producing a new full-length movie based on Aang that will release in January 2026. Early in April the official title of The Legend of Aang: The Last Airbender was announced for this movie.

Friday, April 18th marks 87 years since Superman first showed up in comics pages, and DC is not going to let that slide without at least a little bit of a party.  With James Gunn’s Superman hitting theaters this summer, DC is ramping up for a big few months for Big Blue. We already talked about a bunch of their plans in comics pages – everything from the return of Lex Luthor to Dan Slott’s new Superman book to an all-ages book that I’m almost certainly buying for my kid. But today, on the big Supermanniversary, DC is really going over the top. If you’re in Burbank and you have some time on your hands, there is a sold out, very special, Superman-themed Warner Brothers Studio Tour that starts with a hang with some comic greats – Dustin Nguyen, Jon Bogdanove and Scott Koblish – and ends with a special screening of the classic Donner film, hosted by TCM’s Ben Mankiewicz.  If, however, you are not within spitting distance of southern California nor connected to the seedy underworld of studio tour ticket scalping, DC’s still got you covered. For the school aged among us, they’re running virtual tours of DC’s offices for school groups, with a focus on Superman and storytelling, something that likely pricks up most nerd parents’ ears. And for folks who are too old/cool for school but have a local comic shop available, DC’s got something too: new printings of comic classics.  There are special, Superman Day editions of All-Star Superman and Superman Unchained; and free copies of Superman for All Seasons #1 and All-Star Superman #1 in stores especially for the 18th.  Superman for All Seasons is a classic coming-of-age tale about Clark leaving Smallville, written by Jeph Loeb and drawn by Tim Sale at the peak of their Long Halloween-period. It’s widely regarded as one of the best Superman comics of all time – Sale’s art is phenomenal, especially. This story was the foundation of the Loeb-produced Smallville show, and apparently it’s an influence on the new movie. Superman Unchained was a New 52-era miniseries written by Scott Snyder and drawn by Jim Lee and it’s a very interesting, very well-made comic. It’s interesting because it came from a period in Snyder’s career where he hadn’t yet developed the bombast he would write with post-Metal, but he was also writing for Jim Lee. So, like any sane person, he made the biggest comic he could, and the result was fun and gorgeous.  All-Star Superman is arguably the greatest Superman story ever told. Grant Morrison and Frank Quitely are perfect together and the story is everything right about the character.  Expect plenty more Superman celebrations as the summer goes on, and we’ll have all the news you’ll want to know about it. 

NBA fans are still recovering from the shocking trade of star Luka Doncic from the Mavericks to the LA Lakers. And in a new interview with Doncic, he reveals that his iPhone might still be recovering a little too. NBA star Luka Doncic threw his iPhone after learning of trade, but ‘it still works’ Luka Doncic recently sat down with ESPN’s Malika Andrews to discuss the shocking trade from Dallas to LA. And in the interview, he reveals that when he first got the call with the news, his first reaction was to throw his iPhone across the room. You can find the video here on X or on ESPN’s website. As you can see in the images above and below, the iPhone sustained significant damage. Despite the damage though, Doncic says he’s still using that same iPhone. ”It still works,” he tells Andrews. So what’s the takeaway? Other than the obvious: that Dallas never should have traded Doncic. Perhaps it’s two-fold: The next time you get some bad news, maybe don’t throw your iPhone? But if you do, your phone might be just fine—though who knows Maybe the most surprising detail for me personally? That the aesthetic damage didn’t lead superstar Luka Doncic to buy himself a new iPhone. Best iPhone accessories Add 9to5Mac to your Google News feed.  FTC: We use income earning auto affiliate links. More.You’re reading 9to5Mac — experts who break news about Apple and its surrounding ecosystem, day after day. Be sure to check out our homepage for all the latest news, and follow 9to5Mac on Twitter, Facebook, and LinkedIn to stay in the loop. Don’t know where to start? Check out our exclusive stories, reviews, how-tos, and subscribe to our YouTube channel