Assassins Creed Shadows creative director Jonathan Dumont has revealed that the studio decided to make the open-world in the upcoming title leaner, with less clutter than previous games in the franchise. Speaking to GamesRadar, Dumont spoke about how visiting Japan to do research on Assassins Creed Shadows changed the developers minds on how to tackle things in the game.When you go on location, theres always something that even if you looked at all the books, the movies, everything, you say oh man, thats surprising,' said Dumont. It was the density of the trees and the forest and how much there was. You had a lot of little mountains that created overlapping views that typically we didnt quite get.Dumont talks about the studio needing to focus on crafting a world with an immense level of fidelity even at smaller scales, thanks to having their perspectives changed after seeing a Japanese castle fortress in person.And the scale of castles and things like that. Castles are big, you know, the castle fortresses holy crap, when you go there, its like I didnt expect that,' he said. So we had to adjust to sort of wrap our mind around, well, we needed much more fidelity on the scale.Owing to this, the studio is working on making sure that players will feel a similar level of awe, which can only be achieved when they arent constantly being bombarded by new points of interest.The travel time is a little different its not a point of interest [followed immediately by another] point of interest, [every] 50 meters, explained Dumont. Theres more open, natural landscaping that will be a little bit longer to travel, but when you get to a place theres a lot more to it. The scale ratio changed from Odyssey, and what we had done before, and then just the fidelity of trying to make incredible attention to detail in the environment, so that we craft the crap out of it.After facing quite a few delays, Assassins Creed Shadows finally got a release date of March 20. Game director Charles Benoit revealed that, in this latest delay, the studio isnt working on overhauling any of the games major mechanics. Rather, the studio is spending the extra time to polish whats already in the game.Adding on to this, Benoit also mentioned that a key area of polish in the upcoming title is its parkour system. Since the roofs of the buildings of Feudal Japan tend to be more complex than the places weve previously seen from the franchise, extra work was needed to make sure that protagonists Naoe and Yasuke will be able to run up and down these buildings without looking awkward.Assassins Creed Shadows is coming to PC, PS5 and Xbox Series X/S. Pre-ordering the game also gets players the Claws of Awaji DLC, which will feature an additional 10 hours worth of content, complete with a new zone and new equipment for players to find. The DLC is slated for release later in the year.

With the release of Monster Hunter Wilds being just over the horizon, developers of the title have been revealing more and more details about the upcoming game. In an interview with IGN, director Yuya Tokuda has revealed that the monsters in Wilds will feature larger health pools and higher flinch thresholds than weve seen in World.According to Tokuda, this is because the studio wants players to make frequent use of the games new Focus Strikes mechanic, which essentially allows players to do massive damage to monsters by exploiting the wounds they accrue throughout a fight. Owing to this mechanic, despite monsters having more health, the duration of hunts wont be too different from what players are used to from previous titles.Its not as if that was our intention from the start, but health did end up being a little bit higher than World once we thought about maintaining appropriate playtimes and player satisfaction, compared to past titles, explained Tokuda. Elements like flinch resistance are also higher, but this doesnt mean that hunts will be tedious. Part of the purpose of Focus Mode is to allow players to feel more accomplished through shorter loops. We tried to design it so that the time they spend hunting is even more concentrated than before.Tokuda spoke quite a bit about Focus Strikes, explaining that the studio wanted to make sure that each weapon type feels unique with the mechanic. This led to the studio making unique Focus Strike animations for each weapon. He did, however, admit that Focus Strikes may have been a bit too strong from the beta tests back in November.For Focus Strikes, we wanted to have animations that show each weapons uniqueness, said Tokuda. But I admit that this also created a variation between weapons during the open beta test where some were far too strong, while feeling the strengths of others was difficult. Though we do allow differences between weapons in order to give them personality, we dont want the disparity between weapons to be too extreme, and so were tuning them to be more standardized for the official release of the game.Earlier this month, Tokuda had also spoken about how Capcom decides which old monster to bring back for its new title, and how much of the roster should consist of new monsters. Ultimately, rather than having some sort of quota, the studio instead focuses on what would be an ideal experience for the player, as well as how a monster would fit into the ecology of the world created in the new game.Tokuda also explained that it designed new monsters by taking into account how it would affect the ecology of the zones it is being designed for. Providing the example of Monster Hunter Wilds cover monster Arkveld, Tokuda mentioned that the developers had to consider things like where the monster would sit in the hierarchy of monsters that would be predators and prey, among many other things.Monster Hunter Wilds is coming to PC, PS5 and Xbox Series X/S on February 28.

Earths Magnetic North Pole Is Shifting Toward Siberia and Raising Questions About Unusual MovementScientists released an update to a model that maps the ever-moving pole and has significant implications for navigation systemsElla JeffriesStaff ContributorJanuary 24, 2025 12:41 p.m. The magnetic north pole, distinct from the geographic North Pole, is the point where Earths magnetic field lines converge. NASA via RawpixelFor decades, Earths magnetic north pole has been slowly drifting across the Arctic, but recent shifts in its path have caught the attention of scientistsand those who rely on precise navigation systems. The magnetic north poles movement, which has suddenly accelerated toward Siberia this century, raises questions about whats driving the unusual shift and why its motion matters.The magnetic north pole is distinct from the geographic North Pole, a fixed point where the Earths axis meets its surface. Instead, magnetic north is the changing location where the planets magnetic field lines converge. This field is generated by the churning of the planets outer core, a layer of molten iron and nickel that begins roughly 1,800 miles beneath Earths surface. As the fluid dynamics of the core change, the magnetic pole shifts over time in a process known as the geodynamo. Now, something unusual seems to be happening.The current behavior of magnetic north is something that we have never observed before, William Brown, a global geomagnetic field modeler at the British Geological Survey, says in a statement.Since it was first discovered in 1831 by explorer James Clark Ross, the magnetic north pole has gradually shifted. Over the past century, its movement from Canada toward Russia has accelerated, increasing from about 6 miles per year to a peak of 31 miles annually by the 2000s. However, in the last five years, the rate of movement has slowed significantly to about 22 miles per year, the biggest deceleration in speed weve ever seen, Brown adds in the statement.British and American scientists have traced this motion over time. Every five years, they release a new edition of the World Magnetic Model (WMM), a tool used by GPS systems and navigation devices that maps the north magnetic pole. As of the latest revision, released in December, its location is closer to Siberia than it was five years ago.For industries reliant on magnetic fields, such as aviation, shipping and navigation, this movement is no small matter. GPS systems, planes and military equipment track the magnetic field and rely on accurate models of magnetic north to function properly. When the magnetic field shifts, the models must be updated to reflect the changes.The more you wait to update the model, the larger the error becomes, Arnaud Chulliat, a senior research scientist at the University of Colorado, Boulder, and NOAAs National Centers for Environmental Information, tells Mindy Weisberger of CNN. The way the model is built, our forecast is mostly an extrapolation given our current knowledge of the Earths magnetic field.Scientists are still unsure why the magnetic north poles speed changed so dramatically. Some researchers believe that changes in the strength of the magnetic field near Canada and Siberia could be influencing the shift.Ciarn Beggan, a geophysicist at the British Geological Survey, says the overall strength of Earths magnetic field has been decreasing over the past couple of centuries, but not uniformly. The field is getting weaker in Canada but stronger around Siberia, and that pulls the magnetic pole towards Siberia, Beggan tells Kasha Patel of the Washington Post.However, the exact reasons for the acceleration and subsequent deceleration remain unclear. The behavior of the magnetic field is considered complicated and chaotic, Brown tells Live Sciences Skyler Ware, which is why the WMM must be updated regularly. In 2019, scientists released the WMM one year early, because the magnetic pole was moving much faster than usual.While these changes may seem minor to the average person, they have significant consequences for systems that rely on magnetic fields for navigation. The latest version of the WMM includes a much higher resolution map of magnetic north than ever before to help reduce errors in navigation systems.Scientists say that if you tried traveling the 5,280 miles from South Africa to the United Kingdom in a straight line using the old WMM, it would leave you 93 miles off course by the end of your trip, per the statement.The WMM update also comes amid broader scientific discussions about the future of Earths magnetic field. Over the last two centuries, the field has weakened by about 9 percent, per the Washington Post. Geophysicists have long speculated that the weakening of the magnetic field could signal a potential geomagnetic reversal, in which the planets north and south magnetic poles swap places.While such an event is not expected to happen any time soon, they have occurred periodically throughout Earths historythe last reversal took place around 780,000 years ago. Scientists are not yet able to predict when the next reversal will occur.As for the magnetic north pole, its future movement is also uncertain.It could change (its) rate, or even speed up again, Brown told CNN. We will continue to monitor the field and assess the performance of the WMM, but we do not anticipate needing to release a new model before the planned update in 2030.Get the latest stories in your inbox every weekday.Filed Under: Astrophysics, Canada, Earth's Core, Explorers, Geography, Geology, Navigation, Physics, Russia , Technology

Researchers from the the Polytechnic University of Milan are mappingthe castle's subterranean structures. Polytechnic University of MilanBeneathSforza Castle in Milan, researchers have discovered a series of secret tunnelspassageways once sketched by the Renaissance artist and scientistLeonardo da Vinci and traversed by a widower sick with grief.The tunnels were discovered by experts from the Polytechnic University of Milan, who used ground-penetrating radar and laser scanners to map the 15th-century strongholds subterranean structures. The imaging revealed cavities and passageways positioned just a foot or two below the ground.The aim is to create a digital twin of the Sforza Castle, a digital model that not only shows the current appearance of the castle, but also allows us to explore the past, by revealing ancient structures that are no longer visible, saysFranco Guzzetti, a geomatics researcher at the Polytechnic University of Milan, in astatement. Imaging revealed cavities and passageways positioned just a foot or two below the ground. Polytechnic University of MilanWorkers began building Sforza in 1358, but the castle was destroyed during a period of political unrest about a century later. Francesco Sforza, the duke of Milan, ordered a new castle to be rebuilt on the ruins in 1450. After he died, his son, Ludovicoalso known as il Moro, or theMoor, because of his dark complexion and hairsecured the Duchy of Milan and continued renovating the castle.Ludovico hired several artistsincluding Leonardo, Donato Bramante and Bernardino Zenaleto decorate the structures walls. Leonardos frescoes can be seen in the castles Sala delle Asse, or Room of Wooden Boards. During this time, he also documented secret passages built for military purposes under the castles ghirlanda (outer wall).This passageway is immortalized in Leonardos drawings and has long been the subject of legends and considerable speculation, per the statement. But now, thanks to technology, it seems that its existence can be confirmed. Sforza Castle's construction began in 1358. Graziella taibi via Wikimedia Commons under CC BY-SA 4.0Historical records suggest Ludovico ordered the construction of an underground passageway to allow him easy access to MilansBasilica of Santa Maria delle Grazie, where his wife was buried. Ludovico had married Beatrice dEste, the daughter of the duke of Ferrara, in 1491. As Encyclopedia Britannica writes, their marriage was unusually harmonious but short-lived: Beatrice died in childbirth in 1497. Ludovicos secret tunnelwhich may have stretched over half a milewould have provided the widower private visitations and served as anescape route in case of a siege.The researchers scanned the castles grounds withground-penetrating radar, which works by emitting radio waves into the earth and sending back signals to an antenna. Eventually, these scans could be used to create augmented reality experiences or virtual visits for tourists, as Claudio Salsi, historical director of the castles museums, tells ANSA.Francesca Biolo, an architect at the Polytechnic University of Milan, says in the statement that the scans are only the first step in unraveling the secrets of Sforza Castles tunnels.The [ground-penetrating radar] technology has added to our 3D model by providing data about well-known but inaccessible spaces, revealing unknown pathways and suggesting ideas for further research into these secret passages, she adds.Get the latest stories in your inbox every weekday.

G2 Esports announced the launch of a new media agency called 62, which aims to connect brands with gamers.Read More

Meta CEO Mark Zuckerberg expects to spend as much as $65 billion on AI in 2025 as part of a massive effort to further the companys AI ambitions. Part of the plan includes a Louisiana data center that Zuckerberg says is so large it would cover a significant part of Manhattan, he wrote on Threads today.The announcement reads like a response to the big AI data center news touted by competitors earlier this week.as they announced Project Stargate, a $500 billion joint venture that will build sprawling AI data centers in Texas and other parts of the country. City documents seen by Bloomberg suggest the Texas data center will be as big as New Yorks Central Park.Meta started building its $10 billion AI data center in Richland Parish, Louisiana last December, and its construction is expected to continue through 2030. The data center is one of many Meta started working on last year to power its open-source large language model Llama. Aside from Meta, tech giants including OpenAI, Google, Microsoft, and Nvidia are spending billions to scale up their data-hungry AI efforts.RelatedZuckerberg said he expects to end the year with over 1.3 million GPUs, while significantly growing the companys AI team. This will be a defining year for AI, Zuckerberg wrote. In 2025, I expect Meta AI will be the leading assistant serving more than 1 billion people, Llama 4 will become the leading state of the art model, and well build an AI engineer that will start contributing increasing amounts of code to our R&D [research and development] efforts.

Earlier this week, several posts on the Instagram page of Aid Access, an abortion pill provider, were inaccessible to the public. Some images were blurred out, with no option to click through and view the post. Others appeared simply as a gray square with nondescript alt text, as if the image didnt load. Aid Access connects patients with doctors who provide abortion pills via telehealth appointments, and the posts that were blocked from being viewed included instructions for performing at-home abortions using pills. The issues on Instagram first reported by Jessica Valenti also reportedly made it difficult to find the Aid Access account using the apps search function.By Thursday evening, Meta had restored a handful of Aid Access posts, though some appear to still be missing. This latest incident is just another example of how Meta has restricted abortion information online for years. It also comes in the middle of Metas right-wing pivot, as the company has begun allowing more transphobic, racist, and otherwise hateful content on its platforms as it courts Donald Trump.The Aid Access Instagram account as of Wednesday evening. Image: MetaAn example of how Aid Access posts appeared on Instagram. Image: MetaMedication abortions in the first trimester of pregnancy can safely be done at home, according to the World Health Organization. Licensed providers like those working with Aid Access have prescribed abortion pills to hundreds of thousands of patients. Thanks to shield laws, which protect healthcare workers who provide the procedure, patients in states where abortion is banned or restricted can also order abortion medication.Social media is an important place for patients to seek information about abortions, says Rebecca Davis of Hey Jane, which offers virtual reproductive care like abortions and birth control.[Patients] will often turn to social media to just make sure were legit, says Davis, who leads marketing at the company. Weve spent a lot of time and energy to really build up our social presence, so were verified on Instagram, were verified on TikTok.Hey Jane encountered restrictions on Instagram in recent days as well. Davis says the group has gotten messages that its Instagram profile was not easily accessible through the apps search features. The Verge was able to replicate the issue: typing in heyjane or hey jane did not display the account as suggested. Users would need to know the accounts full handle, @heyjanehealth, in order for it to appear as a suggestion.We know that by not showing up in these searches were directly impacting people who are actively seeking this very timely, essential healthcare from getting the information that they need to make decisions, Davis told The Verge.Are you a current Meta employee with info youd like to share? Contact me at mia@theverge.com or on Signal (@miasato.11) using a non-work device. You can be anonymous.Meta spokesperson Erin Logan told The Verge in an email that abortion rights groups are experiencing a variety of issues some due to correct enforcement, as well as over enforcement. Logan said the company prohibits the sale of pharmaceutical drugs on Meta platforms without a LegitScript certification. (Hey Jane is among the providers certified with LegitScript.) Logan said these incidents were not the result of recent Meta policy updates.Weve been quite clear in recent weeks that we want to allow more speech and reduce enforcement mistakes and were committed to doing that, Logan added, though no specifics were mentioned. Regarding issues encountered by Hey Jane, Logan directed The Verge to Metas policies, which state that content promoting the use of pharmaceutical drugs is allowed on the platform but may not be eligible for recommendation. In recent weeks, Meta has rolled back its policies for what users can and cant say on its platforms, opening the floodgates for more hate speech and offensive content under the guise of free expression. But providers have long had information about abortion restricted or removed, according to groups like Amnesty International and Repro Uncensored. Davis from Hey Jane says this isnt the first time the groups Instagram profile has been invisible in search: something similar happened in 2023, when fake Hey Jane accounts were appearing in search instead.Even though abortion access groups have encountered issues on Meta platforms, Davis says its not as simple as moving to another social media site. Many patients use Instagram and other platforms like TikTok to find urgent information.While this moment certainly points to the value in diversifying the platforms that were on, that doesnt necessarily mean that people who are seeking abortion care are going to be off of these platforms, Davis says.

By intertwining the development of artificial intelligence combined with large language models with reinforcement learning in high-performance computation, the newly developed Reasoning Language Models may leap beyond traditional ways of limitation applied to processing by language systems toward explicit and even structured mechanisms, enabling complex reasoning solutions across diverse realms. Such model development achievement is the next significant landmark for better contextual insights and decisions.The design and deployment of modern RLMs pose a lot of challenges. They are expensive to develop, have proprietary restrictions, and have complex architectures that limit their access. Moreover, the technical obscurity of their operations creates a barrier for organizations and researchers to tap into these technologies. The lack of affordable and scalable solutions exacerbates the gap between entities with access to cutting-edge models, limiting opportunities for broader innovation and application.Current RLM implementations rely on complex methodologies to achieve their reasoning capabilities. Techniques like Monte Carlo Tree Search (MCTS), Beam Search, and reinforcement learning concepts like process-based and outcome-based supervision have been employed. However, these methods demand advanced expertise and resources, restricting their utility for smaller institutions. While LLMs like OpenAIs o1 and o3 provide foundational capabilities, their integration with explicit reasoning frameworks remains limited, leaving the potential for broader implementation untapped.Researchers from ETH Zurich, BASF SE, Cledar, and Cyfronet AGH introduced a comprehensive blueprint to streamline the design and development of RLMs. This modular framework unifies diverse reasoning structures, including chains, trees, and graphs, allowing for flexible and efficient experimentation. The blueprints core innovation lies in integrating reinforcement learning principles with hierarchical reasoning strategies, enabling scalable and cost-effective model construction. As part of this work, the team developed the x1 framework, a practical implementation tool for researchers and organizations to prototype RLMs rapidly.The blueprint organizes the construction of RLM into a clear set of components: reasoning schemes, operators, and pipelines. Reasoning schemes define the structures and strategies for navigating complex problems ranging from sequential chains to multi-level hierarchical graphs. Operators control how these patterns change so that operations can smoothly include fine-tuning, pruning, and restructurings of reasoning paths. Pipelines allow easy flow between training, inference, and data generation and are adaptable across applications. This block-component structure supports individual access while models can be fine-tuned to a fine-grained task such as token-level reasoning or broader structured challenges.The team showcased the effectiveness of the blueprint and x1 framework using empirical study and real-world implementations. This modular design provided multi-phase training strategies that could optimize policy and value models, further improving reasoning accuracy and scalability. It leveraged familiar training distributions to maintain high precision across applications. Noteworthy results included large efficiency improvements in reasoning tasks attributed to the streamlined integration of reasoning structures. For instance, it demonstrated the potential for effective retrieval-augmented generation techniques through experiments, lowering the computational cost of complex decision-making scenarios. Such breakthroughs reveal that the blueprint allows advanced reasoning technologies to be democratized to even low-resource organizations.This work marks a turning point in the design of RLMs. This research addresses important issues in access and scalability to allow researchers and organizations to develop novel reasoning paradigms. The modular design encourages experimentation and adaptation, helping bridge the divide between proprietary systems and open innovation. The introduction of the x1 framework further underscores this effort by providing a practical tool for developing and deploying scalable RLMs. This work offers a roadmap for advancing intelligent systems, ensuring that the benefits of advanced reasoning models can be widely shared across industries and disciplines.Check out the Paper. All credit for this research goes to the researchers of this project. Also,dont forget to follow us onTwitter and join ourTelegram Channel andLinkedIn Group. Dont Forget to join our70k+ ML SubReddit. Nikhil+ postsNikhil is an intern consultant at Marktechpost. He is pursuing an integrated dual degree in Materials at the Indian Institute of Technology, Kharagpur. Nikhil is an AI/ML enthusiast who is always researching applications in fields like biomaterials and biomedical science. With a strong background in Material Science, he is exploring new advancements and creating opportunities to contribute. Meet 'Height':The only autonomous project management tool (Sponsored)

The Steepest Ascent Hill Climbing Algorithm Unraveled 0 like January 24, 2025Share this postAuthor(s): Mirko Peters Originally published on Towards AI. The steepest ascent hill climbing algorithm is a potent tool for optimization, yet it faces challenges such as local maxima, evaluation function dependency, and plateaus. By overcoming these hurdles through clever adaptations and real-world applications, we can harness its potential effectively.This member-only story is on us. Upgrade to access all of Medium.Source: Data & AnalyticsImagine scaling a mountain, wind in your hair, feeling invincible. But as you reach the summit, you realize theres an even taller peak just over the horizon. This metaphor perfectly captures the pursuit of optimization in algorithms, particularly in steepest ascent hill climbing. Join me as I delve into this powerful but nuanced algorithm, its challenges, and its real-world impact, enriched with victory tales and pitfalls from my own experiences. The Hero and the Villains: Introduction to Steepest Ascent Hill Climbing The Local Maxima Dilemma The Pitfalls of Poor Evaluation Functions Facing Flat Regions and Plateaus Innovations and Solutions in Optimization Real-World Applications of Steepest Ascent Hill Climbing The Future: Building Smarter AlgorithmsHow to optimize the steepest ascent hill climbing algorithm?When I first dove into the steepest ascent hill climbing algorithm, I was struck by how similar it is to climbing a physical hill. Youre trying to reach the top, but the path isnt always clear. The algorithm works by evaluating neighboring solutions and moving toward the one that offers the best improvement. Simple, right? But its not just a linear climb; its a journey filled 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 asponsor. Published via Towards AITowards AI - Medium Share this post

GenAI for Better NLP Systems II: Social Media Analytics 0 like January 24, 2025Share this postLast Updated on January 24, 2025 by Editorial TeamAuthor(s): Nabanita Roy Originally published on Towards AI. Leveraging GenAI for Social Media Analytics for swifter text processing and inferences as well as improving traditional NLP models with the aid of LLMsThis member-only story is on us. Upgrade to access all of Medium.Photo by Adem AY on UnsplashGenerative AI, especially OpenAIs ChatGPT, has revolutionized how we approach data. Leveraging this modern tool (or should I call it futuristic), we can overcome many challenges in data-driven systems if used intelligibly. In my previous article, I demonstrated how we can use GenAI to generate synthetic data to boost performance for machine learning models for unbalanced data for Natural Language Processing (NLP) use-cases. This article shifts focus to social media analytics, exploring how GenAI can:Accelerate NLP inferencesEnhance datasets for traditional modelsOvercome limitations of traditional NLP techniques with easeHere, I have expanded on the most common NLP tasks in social media analytics and demonstrated the usage of prompting techniques to get LLMs to do the job. However, Large Language Models (LLMs) can either be used to solve the problem end-to-end or can be used for auxiliary tasks such as data labeling, feature extraction, or preliminary analysis, thus enhancing the process of building traditional models.In each segment of this article, Ive included a How to Use section that highlights how tasks performed using LLMs can be leveraged to support the development of traditional NLP models. 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 asponsor. Published via Towards AITowards AI - Medium Share this post