FlexiSpot has produced another WFH game-changer.

Sonys recent State of Play saw some mixed opinions from players, especially with the relative lack of PlayStation Studios titles (barring Housemarques Saros). However, it also brought up the relative lack of news for Fairgame$. Announced in May 2023, Haven Studios co-op heist title has received no updates since (barring a job listing).There doesnt seem to be any hope for a shadow drop this year. On the recent Game Mess Decides, Giant Bombs Jeff Grubb revealed, No Fairgame$ this year. Fairgame$, as far as Im aware, has been pushed to 2026.Its somewhat understandable the staggering failure of Concord reportedly led to Sony reconsidering its live service plans and cancelling projects from Bend Studio and Bluepoint Games. Perhaps its taking more time for polish to ensure a satisfactory experience. Maybe there are some overhauls. Either way, it may be a while before its playable.The premise of Fairgame$ sees players as heisters who venture into forbidden locations worldwide to steal cool loot. Unraveling the nefarious plans of billionaires is also on the menu, as the gameplay focuses on emergent sandbox PvP. Its currently being developed for PS5 and PC. Stay tuned for more updates in the coming months.

Alongside Haven Studios Fairgame$, Bungies extraction shooter Marathon was also missing during Sonys latest State of Play. The latter noted that playtests would expand this year and even released character art for two Runners codenames Thief and Stealth but when can fans expect more information?According to Giant Bombs Jeff Grubb, April may be the month. Speaking on the latest Game Mess Decides, he reported that Marathon was absent from State of Play since Theyre probably going to do their own thing.Bungie does their own thing usually for Destiny, theyre going to adopt that same Were talking to our audience strategy this year as well, so I think April. Well probably hear from them about Marathon, and it would just be its own thing separate from a State of Play. How concrete these plans are and how much will be revealed is unknown.However, if Bungie wants to get ahead of the next rumored PlayStation Showcase while leaving room for Destiny 2s future content reveal (with Codename Apollo out in Summer), April would be fitting. Time will tell, so stay tuned for updates.Last year saw rumors circulating about the not great sentiment around Marathon and how it could miss its internal 2025 release target. Its slated for Xbox Series X/S, PS5, and PC.

In this tutorial, well learn how to create a custom tokenizer using the tiktoken library. The process involves loading a pre-trained tokenizer model, defining both base and special tokens, initializing the tokenizer with a specific regular expression for token splitting, and testing its functionality by encoding and decoding some sample text. This setup is essential for NLP tasks requiring precise control over text tokenization.from pathlib import Pathimport tiktokenfrom tiktoken.load import load_tiktoken_bpeimport jsonHere, we import several libraries essential for text processing and machine learning. It uses Path from pathlib for easy file path management, while tiktoken and load_tiktoken_bpe facilitate loading and working with a Byte Pair Encoding tokenizer.tokenizer_path = "./content/tokenizer.model"num_reserved_special_tokens = 256mergeable_ranks = load_tiktoken_bpe(tokenizer_path)num_base_tokens = len(mergeable_ranks)special_tokens = [ "<|begin_of_text|>", "<|end_of_text|>", "<|reserved_special_token_0|>", "<|reserved_special_token_1|>", "<|finetune_right_pad_id|>", "<|step_id|>", "<|start_header_id|>", "<|end_header_id|>", "<|eom_id|>", "<|eot_id|>", "<|python_tag|>",]Here, we set the path to the tokenizer model, specifying 256 reserved special tokens. It then loads the mergeable ranks, which form the base vocabulary, calculates the number of base tokens, and defines a list of special tokens for marking text boundaries and other reserved purposes.reserved_tokens = [ f"<|reserved_special_token_{2 + i}|>" for i in range(num_reserved_special_tokens - len(special_tokens))]special_tokens = special_tokens + reserved_tokenstokenizer = tiktoken.Encoding( name=Path(tokenizer_path).name, pat_str=r"(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^rnp{L}p{N}]?p{L}+|p{N}{1,3}| ?[^sp{L}p{N}]+[rn]*|s*[rn]+|s+(?!S)|s+", mergeable_ranks=mergeable_ranks, special_tokens={token: len(mergeable_ranks) + i for i, token in enumerate(special_tokens)},)Now, we dynamically create additional reserved tokens to reach 256, then append them to the predefined special tokens list. It initializes the tokenizer using tiktoken. Encoding with a specified regular expression for splitting text, the loaded mergeable ranks as the base vocabulary, and mapping special tokens to unique token IDs.#-------------------------------------------------------------------------# Test the tokenizer with a sample text#-------------------------------------------------------------------------sample_text = "Hello, this is a test of the updated tokenizer!"encoded = tokenizer.encode(sample_text)decoded = tokenizer.decode(encoded)print("Sample Text:", sample_text)print("Encoded Tokens:", encoded)print("Decoded Text:", decoded)We test the tokenizer by encoding a sample text into token IDs and then decoding those IDs back into text. It prints the original text, the encoded tokens, and the decoded text to confirm that the tokenizer works correctly.tokenizer.encode("Hey")Here, we encode the string Hey into its corresponding token IDs using the tokenizers encoding method.In conclusion, following this tutorial will teach you how to set up a custom BPE tokenizer using the TikToken library. You saw how to load a pre-trained tokenizer model, define both base and special tokens, and initialize the tokenizer with a specific regular expression for token splitting. Finally, you verified the tokenizers functionality by encoding and decoding sample text. This setup is a fundamental step for any NLP project that requires customized text processing and tokenization.Here is the Colab Notebook for the above project. Also,dont forget to follow us onTwitterand join ourTelegram ChannelandLinkedIn Group. Dont Forget to join our75k+ 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/LG AI Research Releases NEXUS: An Advanced System Integrating Agent AI System and Data Compliance Standards to Address Legal Concerns in AI DatasetsAsif Razzaqhttps://www.marktechpost.com/author/6flvq/KAIST and DeepAuto AI Researchers Propose InfiniteHiP: A Game-Changing Long-Context LLM Framework for 3M-Token Inference on a Single GPUAsif Razzaqhttps://www.marktechpost.com/author/6flvq/DeepSeek AI Introduces CODEI/O: A Novel Approach that Transforms Code-based Reasoning Patterns into Natural Language Formats to Enhance LLMs Reasoning CapabilitiesAsif Razzaqhttps://www.marktechpost.com/author/6flvq/Google DeepMind Researchers Propose Matryoshka Quantization: A Technique to Enhance Deep Learning Efficiency by Optimizing Multi-Precision Models without Sacrificing Accuracy [Recommended] Join Our Telegram Channel

Understandability of Deep Learning Models 0 like February 15, 2025Share this postLast Updated on February 17, 2025 by Editorial TeamAuthor(s): Lalit Kumar Originally published on Towards AI. This member-only story is on us. Upgrade to access all of Medium.Blackbox nature of DL modelsDeep learning systems are a kind of black box when it comes to analysing how they give a particular output, and as the size of the model increase this complexity is further increased. These models despite their impressive performance across various domains, often suffer from lack of transparency issue. Their internal workings are very complex and not easy to understand, hence they are sometimes also referred as black boxes. This lack of transparency hinders trust and limits their applicability in safety-critical domains. It is difficult to judge how these powerful models arrive at their decisions. This challenge, often referred to as the deep learning understandability problem, has spurred significant research efforts to develop techniques that shed light on the inner workings of these models. For, a smaller model, it may be possible to explore the internal representations and try to understand the model's decision-making process. But as the model size increases so is the problem to understand its decision-making mechanism.Then, how to keep a track of these models functioning and interpret them?Following are some of the solutions which handle Deep Learning Models understandability problem:This technique 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

LatestMachine LearningA Unified Approach To Multimodal Learning 0 like February 14, 2025Share this postLast Updated on February 17, 2025 by Editorial TeamAuthor(s): Yash Thube Originally published on Towards AI. A Unified Approach To Multimodal LearningWe live in a world of multimodal data. Think about it: a restaurant review isnt just text; its accompanied by images of the food, the ambiance, and maybe even the menu. This combination of text, images, and ratings is multimodal data, and its everywhere. Traditional machine learning models often struggle with this kind of data because each mode (text, image, rating) has its unique structure and characteristics. A picture is high-dimensional, text is sequential, and a rating is just a number. How do we effectively combine these disparate data types to make better predictions?The paper Generative Distribution Prediction: A Unified Approach to Multimodal Learning introduces a new framework called Generative Distribution Prediction (GDP) to tackle this challenge. GDPs core idea is to use generative models to understand the underlying process that creates this multimodal data. Imagine an artist trying to recreate a scene. They dont just copy it; they understand the relationships between the objects, the lighting, and the overall composition. Similarly, generative models learn the underlying structure of the data, allowing them to create synthetic data that resembles the real thing. This synthetic data, as well see, is key to improving predictions.The Problem: Multimodal Data is MessyIt offers a clever solution: it uses generative models to create synthetic data that captures the combined information from all modalities. Think of it as the model learning to imagine new restaurant reviews, complete with pictures and ratings, based on what it has already seen. By learning this generative process, the model gains a deeper understanding of the relationships between the different modes. This understanding then allows it to make better predictions on real data.How GDP Works: Two-Step ProcessIt works in two main steps:Constructing a Conditional Generator: This step focuses on building a generative model that can create synthetic data conditioned on specific input values. For example, the model might generate a synthetic restaurant review (text, image, rating) given a specific cuisine type and price range. This often involves transfer learning, where a pre-trained generative model is fine-tuned on the specific multimodal data. A key component here is the use of dual-level shared embeddings (DSE). Embeddings are a way of representing data as vectors of numbers, capturing semantic meaning. DSE creates shared embeddings at two levels, helping the model to learn relationships between different modalities and also adapt to new, unseen data (a process called domain adaptation).Using Synthetic Data for Point Prediction: Once the conditional generator is trained, it can be used to create synthetic data for any given input. This synthetic data represents the possible responses associated with that input. The model then makes a prediction by finding the response that minimizes the prediction error on this synthetic data. This is like the model saying, Based on what Ive learned about how reviews are generated, this is the most likely rating for this restaurant.Why is it Better?Unified Framework: It handles multimodal data within a single generative modeling framework, eliminating the need for separate models for each modality.Mixed Data Types: It can handle different data types (text, images, tabular data) seamlessly, modeling the conditional distribution of the variables of interest.Robustness and Generalizability: By training on synthetic data, GDP becomes more robust to noise and variations in the real data, improving its ability to generalize to new, unseen examples.Key Contributions and Theoretical FoundationsThe paper makes several important contributions:GDP Framework: Introduces the GDP framework for multimodal supervised learning using generative models.Theoretical Foundation: Provides theoretical guarantees for GDPs predictive accuracy, especially when using diffusion models as the generative backbone. It analyzes two key factors: generation error (how different the synthetic data is from the real data) and synthetic sampling error (the error introduced by using a finite sample of synthetic data).Domain Adaptation: Proposes a novel domain adaptation strategy using DSE to bridge the gap between different data distributions.Multimodal Diffusion Models: The Generative EngineA crucial component of GDP is the use of diffusion models as the generative engine. Diffusion models are a powerful type of generative model that works by gradually adding noise to data until it becomes pure noise, and then learning to reverse this process to generate data from noise. The paper introduces a specialized diffusion model for multimodal data, integrating structured tabular data with unstructured data like text and images through shared embeddings and a shared encoder-decoder architecture.Numerical Examples and ResultsThe paper evaluates GDP on a variety of tasks, including:Domain adaptation for Yelp reviewsImage captioningQuestion answeringAdaptive quantile regressionThe results consistently show that GDP outperforms traditional models and state-of-the-art methods in terms of predictive accuracy, robustness, and adaptability.In simple terms, It is Like a master chef. Imagine a master chef who has tasted thousands of dishes. They dont just memorize the recipes; they understand the complex interplay of flavors, textures, and ingredients. GDP is like that chef. It learns the underlying recipe for multimodal data, allowing it to generate new dishes (synthetic data) and, more importantly, make better predictions about the real dishes it encounters. By understanding the generative process it unlocks a reasonable potential of multimodal data, leading to more accurate and robust predictions across a wide range of applications.The future directions involve making GDP more computationally efficient, applying it to a wider range of problems, and developing a deeper theoretical understanding of its properties with various generative models.Stay Curious.See you in the next one!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

Full spoilers follow for The White Lotus through Season 3, Episode 1.While the new season of Mike Whites hit HBO show The White Lotus has of course given us a bunch of new characters to welcome, inspect, and judge, it also has seen the return of not one but two faces from past seasons. We already knew that Natasha Rothwells Belinda Lindsey would be back from Season 1, but in a surprise moment, Jon Gries Greg Hunt, also from Season 1 as well as Season 2, also appeared in this first episode. And you know we all hate GregLets brush up on our Greg history, why hes public enemy number one for White Lotus fans, and what his reappearance in the premiere might mean for the new season.Jon Gries as Greg Hunt in The White Lotus.Greg, Belinda, and Tanya in The White Lotus Season 1If The White Lotus ensemble casts have had a star player in the past, that honor would go to Jennifer Coolidges Tanya McQuoid, the hilarious but sort of lost soul of the first two seasons. As with many of the clients of The White Lotus resort, Tanya was very wealthy but also troubled. When we first met her in Season 1, she was visiting the Maui White Lotus with the intention of spreading her recently passed mothers ashes. Spa manager Belinda helped Tanya through her grief and anxiety, so much so that Tanya proposed financing Belinda so that she could start her own spa.Tanya also met Greg here, and a romantic relationship began between the two. But Tanya, suffering from constant anxiety and self-doubt, was trepidatious about it. As she said to Belinda, I just know Im gonna get hurt. Tanya tried to end the relationship with Greg as a result, but his apparent good cheer and compassion for Tanya assuaged her fears. (Wed eventually learn that she shouldve trusted her instincts.)Eventually, Greg revealed to Tanya that he was suffering from a serious illness and that she shouldnt be surprised if he dropped dead at any time. Deciding to leave with Greg for Honolulu, and then possibly move to Aspen to be closer to him, Tanya also opted to not go into business with Belinda. Belinda was crushed by this decision but perhaps not surprised in the end.The Death of Tanya in Season 2Tanya returned for Season 2, as did Greg for the first few episodes, when the pair vacationed at the White Lotus in Sicily. Now married, with Gregs health issues resolved thanks to Tanyas wealth, things have changed. The type of toxic relationship of her past that she worried about in Season 1 seems to have taken root with Greg, and his previously amorous attitude to Tanya has dissipated when she prompts him for some action in the first episode of the season, he announces that he has to go wash up first because he has swamp crotch, for example. Romantico!Later, Tanya hears Greg on the phone in the bathroom whispering. He claims its just a work thing, and she brushes it off. But then in Episode 2, when she wakes up in bed, she finds Greg just blankly staring into the middle distance. Its an odd moment, but it passes as he claims hes fine. At breakfast, Tanya describes her idea of a perfect day in Italy, and Greg, acting nicer than he did the day before, agrees to the plan. Its your day to shine, he says. But that night he tells Tanya that he has to leave their vacation early for work and return to the U.S. This leads to an argument where he reminds her that she made him sign a prenup and that he cant know that she wont discard him at some point the way she has many of her friends and employees over the years. The episode ends with Greg sneaking a phone call where he tells the person on the other end of the line that he loves them and that Tanya is clueless as usual.Greg apparently had plotted with Quentin to kill Tanya and make it look like an accident so that he could inherit her fortune. After Greg leaves, Tanya is befriended by a group of men who she eventually begins to suspect are out to kill her. By the Season 2 finale, Tanyas suspicions are proven correct, but she takes out her would-be assassins with a gun in a glorious moment of haphazard triumph, only to then awkwardly die herself when attempting to get off the yacht where the final showdown took place.The thing is, before she died, Tanya saw a picture of a young Greg with Quentin (Tom Hollander), one of the men who was trying to kill her. Greg apparently had plotted with Quentin to kill Tanya and make it look like an accident so that he could inherit her fortune. Whether or not the authorities ever figured out this plan is unknown by the end of Season 2.The White Lotus Season 3 GalleryWhat Is Greg Doing in Season 3?Good question! We dont know if he actually did inherit Tanyas money or not. Its possible that the police never connected him to the plot which resulted in multiple deaths, including Tanyas. What we do know is that he is now hanging around the Thai White Lotus where Season 3 is set, and he apparently has a younger ex-model for a girlfriend now.Chelsea (Aimee Lou Wood), who is the girlfriend of Walton Goggins Rick Hatchett, meets Gregs girlfriend at the bar, and they bond over their boyfriends who are both older, bald or balding, and generally cranky. Greg also apparently now lives somewhere near the Thai White Lotus.But does he really? Is it possible this is another scam of Gregs, and that Chelsea is being sucked into something she doesnt understand? Will Belinda, back for Season 3, manage to find justice for Tanya by somehow outing Greg as being responsible for her death? Theres no reason for us to think she suspects Greg of any wrongdoing (and she hasnt even encountered him yet at the resort), plus she doesnt necessarily owe Tanya anything after the way she was treated. But Belinda is also one of the good figures in The White Lotus that we are always rooting for, and it seems like if anyone can right the wrongs that were done to Tanya, its her. Of course, ultimately, one of the big questions The White Lotus is always asking is whether or not good people ever really win in the end. Belinda, were counting on you.Will Natasha Rothwells Belinda Lindsey bring justice for Tanya?

Apple will debut its first new product of the year this Wednesday. Apple CEO Tim Cook announced the launch through a tweet last week. While we cant say for certain that the new product launch will be, its highly likely to be the new iPhone SE 4, which may receive a new name.New iPhone SE: most likelyGiven all previous reporting, the most likely outcome from Wednesdays announcement is to be a new iPhone SE model.To quickly recap, were expecting a new 6.1-inch OLED display, an upgrade to USB-C, a 48MP rear camera, Face ID, as well as many other modern iPhone features. The new iPhone SE will also be the testbed for Apples first in-house 5G modem.Some sketchier rumors suggested that Apple may choose to brand this new iPhone SE model with a new moniker: iPhone 16E. Tim Cook referred to the new product launch as the newest member of the family. While this doesnt say anything for certain, many are speculating that it likely suggests that the iPhone 16E naming will come to fruition. Only time will tell.Other plausible announcementsSome other products that are on track to launch early this year include:Refreshed MacBook Air with M4 chipiPad 11th gen with A17 Pro chip (Apple Intelligence update)iPad Air M3 with new Magic KeyboardThese are more than likely launching later in the spring, but they are technically on the table for this weeks Apple announcement.What not to expectApple does have a number of exciting announcements in the pipeline for the first half of the year, but many wont necessarily fall within early spring.For example, AirTag 2 shouldnt launch until mid 2025, according to Bloombergs Mark Gurman. This AirTag upgrade will include a couple of nice quality of life upgrades, including much improved precision finding.Apple is also working on a new smart home hub, which some refer to as HomePad. While this product was initially on track for a March debut, recent reports suggest that it might not hit the market until a little bit later.Of course, theres also M4 Ultra Mac Studio and Mac Pro in the works, which wont launch until WWDC this summer.My favorite Apple accessories on Amazon:Follow Michael:X/Twitter,Bluesky,InstagramAdd 9to5Mac to your Google News feed. FTC: We use income earning auto affiliate links. More.Youre 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. Dont know where to start? Check out our exclusive stories, reviews, how-tos, and subscribe to our YouTube channel

Ikon Gallery will presentWhat Did You Want to See?, a solo exhibition by British artist Mahtab HussainBirmingham Central MosqueSource: Mahtab HussainAl Masjid Al SaifeeSource: Mahtab HussainChasma E-RahmatSource: Mahtab Hussain1/3show captionBirminghams Ikon Gallery is to host What Did You Want to See?, a solo exhibition by British artist Mahtab Hussain, from 20 March to 1 June 2025. Commissioned by Ikon and Photoworks, the exhibition will examine Birminghams built environment, focusing on the citys mosques and urban spaces while addressing the wider context of surveillance and representation.A large-scale photographic installation will systematically document 160 mosques across Birmingham.The collection highlights the varied forms of mosque architecture in the city, from Birmingham Central Mosques domes and minarets to converted terraced houses and repurposed churches.Alongside the mosque studies, the exhibition will feature black and white portraits of Birmingham residents taken in 2024, depicting members of the citys Muslim community in a range of settings.Source: Mahtab HussainRaza MosquePart of the exhibition will reference surveillance culture, with an interactive element designed to evoke the experience of being observed.Other interventions in the gallery space will engage with Birminghams urban landscape, including graffiti-style postcode tagging as a marker of community identity and remnants of Project Championa controversial 2011 surveillance initiative.Tarmac patches, left behind where security cameras were removed, will be repurposed as a visual reference to past monitoring of public spaces.The exhibition will run at Ikon Gallery, Birmingham, from 20 March to 1 June 2025. Admission is free, with donations welcome.>> Also read:Art show explores modern architectures links with horror

Source: Dematerial5plus Architects has secured planning approval from Merton Council for Wimbledon Works, a 12-dwelling residential scheme designed for Goldcrest Land. The project regenerates an underutilised site through the introduction of a mews typology, comprising two linked terraces of four-storey townhouses.The scheme is made up of narrow-fronted, two-bedroom dwellings, each incorporating a separate work-from-home space at ground floor level. The homes are designed with private roof gardens and enclosed winter gardens at roof level.One unit within the development is proposed as a unique, oversized dwelling, designed to be fully adaptable with provision for an internal lift in line with London Plan guidelines. This approach is intended as an alternative to the standard provision of accessible housing within single-floor flats, offering a townhouse typology that accommodates a range of accessibility needs.Source: DematerialThe material palette has been selected to reference the surrounding area and contribute to what the architects describe as a unique sense of place.Only two car parking spaces are proposed, both designated for Blue Badge permit holders, with additional provision for communal bikes.The project has a construction value of 5 million and a total floor area of 15,000 sq ft.Project TeamBuilding Services Engineer: EnsphereSustainability Consultant: EnsphereLandscape Architect: SpacehubPlanning Consultant: RPSFire Engineer: SocotecAcoustic Consultant: KP AccousticsTransport: Vectos