This was my entry for the Pwnishers Chasms Call challenge! It was a bit hard to find some free time to finish it between work, house renovation and game developing but here it is! :sweat_smile: All done i

ServiceNow has upgraded its AI agent offerings as it seeks it sto double down on investing in AI for enterprises.Read More

After embracing generative AI as a core part of its gaming strategy, Netflix's vice president of generative AI has left the company.Read More

Zenless Zone Zero voice actors quietly recast following SAG-AFTRA action"I stand by my choice," actor says News by Vikki Blake Contributor Published on March 12, 2025 Two Zenless Zone Zero voice actors claim they have been replaced for participating in the SAG-AFTRA strikes.As spotted by Eurogamer, Emeri Chase and Nicholas Thurkettle - playing Soldier 11 and Lycaon, respectively - claim they found out their voices had been replaced by reading the patch notes at the same time players did.As it began development before the strike was called on July 25, 2024, Zenless Zone Zero is not subject to the strike, however, voice actors both in and out of the union can choose to support the action.In a Bluesky thread, Chase said: "I was replaced as Soldier 11 because I am unwilling to perform work not covered by a SAG Interim Agreement during a strike for AI protection, the outcome of which will determine the future of our industry."I'd like to clarify that there's a difference between being 'struck' and not being on an Interim Agreement. Union projects that began work prior to the strike and non-union projects are not 'struck.' But they also do not offer the Union-enforced AI rights we are fighting for."Many actors are choosing to voluntarily withhold work on these categories of projects because we feel it is the best way to support the Union's fight for the protections that are critical to our continued ability to create the art we love. I knew that by withholding work it was possible I'd be replaced, though of course I hoped they would choose to leave her silent until I was able to return. I found out the role was recast today alongside all of you."Chase closed on saying they had loved working on the game but [stood] by [their] choices "regardless of the outcome."Thurkettle similarly discussed the issue on Bluesky, saying: "This is what I've been quiet about. I'm not SAG but what game companies want to do with AI is an existential threat. I took a personal stand to ask for protection, and had to be willing to give up the best thing that's ever happened in my professional life. I stand by my choice."GamesIndustry.biz has reached out to YoHoVerse for comment.It follows news earlier this month that the voice talent team that brings Apex Legends' characters to life in French refused to sign contracts that would authorize their voices to be used to train AI.SAG-AFTRA voted to strike at the end of July after it failed to reach an agreement with the convenience bargaining group over rights and protection concerns raised by the industry's exploration of AI technologies.SAG-AFTRA secured agreements with the developers behind 80 upcoming games in September as it staged strike action over better conditions for actors, including protection against the use of AI. In November, SAG-AFTRA announced a new agreement for video game localisation.

Throughout his voice acting career, Mallow voiced characters in Power Rangers and Digimon, and in games like Street Fighter, Resistance 3, and Call of Duty.

Lee previously worked on Final Fantasy IX and Wing Commander III before taking on managerial and producer positions at Nexon America and Stealth CrossMedia.

Meta has spent more than a year advocating for new laws requiring app stores to give parents control over kids app downloads, and just saw an early victory in the states. But Google charges that its really just a misguided effort to offload Metas own responsibility to keep kids safe.The missive follows the passage of Utahs App Store Accountability Act, the first of its kind to advance to the governors desk, putting the onus on app store operators to keep kids from accessing inappropriate content. There are similar bills in more than a dozen states across the country in a growing trend of kids safety legislation, in the wake of the Kids Online Safety Acts failure to become law last year, and ongoing legal battles over many other state laws.While Meta, Snap, and X issued a joint statement praising the Utah bills passage, Google calls it concerning. Rather than protect kids and give parents more control, Google director of public policy Kareem Ghanem writes, the legislation requires app stores to share if a user is a kid or teenager with all app developers (effectively millions of individual companies) without parental consent or rules on how the information is used. That raises real privacy and safety risks, like the potential for bad actors to sell the data or use it for other nefarious purposes. Social media companies would be the real beneficiaries of the law, Ghanem writes, because they could avoid that responsibility despite the fact that apps are just one of many ways that kids can access these platforms. Both Meta and Googles YouTube have come under fire in the past for allegedly not doing enough to keep its youngest users safe on their platforms by pushing videos of kids to potential predators or keeping teens in a content loop that makes them feel bad about themselves. Both companies have said they maintain robust policies and resources to create healthy experiences on their platforms.We welcome Googles concession that they can share age information with app developers, and we agree this should be done in a privacy-preserving manner, Meta spokesperson Jamie Radice says in a statement. But with millions of apps on Googles app store, and more added every day, its unclear how theyll determine which apps are eligible to receive this data. The simplest way to protect teens online is to put parents in charge. Thats why legislation should require app stores to obtain parental consent before allowing children to download apps. In the past, Meta has argued that the app store is the optimal place for parents to grant permission and to vet users ages before they ever download apps. This method would also protect users privacy, Meta global head of safety Antigone Davis wrote in 2023, because by verifying a teens age on the app store, individual apps would not be required to collect potentially sensitive identifying information. How exactly users ages get verified is a major concern for privacy advocates, but its one thats not yet entirely worked out in some of the legislation. Utahs, for example, says that app store operators can use either commercially available methods that are reasonably designed to ensure accuracy, or other methods to be determined and deemed acceptable by state regulators.Because developers know their apps best, they are best positioned to determine when and where an age-gate might be beneficial to their usersGoogle believes it has a better way. To Google, that means that app stores should only provide age assurance securely to developers that actually need them meaning only for apps that offer risky content, and probably not for something more mundane like a weather app. In that vein, Google proposes putting more discretion on app developers, rather than app stores, to determine the appropriate protections to put in place for a given age group. Because developers know their apps best, they are best positioned to determine when and where an age-gate might be beneficial to their users, and that may evolve over time, which is another reason why a one-size-fits-all approach wont adequately protect kids, Ghanem writes. Google is also proposing clear consequences for developers who violate users trust by doing things like improperly accessing or sharing the age signal.Apple has similarly raised concerns about potentially excessive data collection. In a white paper announcing steps it would take to help protect kids online, including letting parents share kids age ranges with developers, Apple emphasized the importance of collecting just the minimal amount of data to protect users privacy.Everyone wants to protect kids and teens online, and make sure they engage with age-appropriate content, Ghanem writes, but how its done matters.See More:

The United Kingdoms mobile browser market is not working well for consumers and businesses according to a final report from Britains competition watchdog, which says that Apple and Google are largely to blame.An independent inquiry group has concluded its mobile browsers investigation for the Competition and Markets Authority (CMA), identifying Apples policies around iOS, Safari, and WebKit as making it difficult for third-party web browser providers to compete and restricting the market as a result. Googles Android mobile ecosystem is also complicit in impeding competition, according to the CMA report, albeit to a lesser extent.The issues flagged by the investigation include Apple requiring all browsers on iOS to run on its WebKit browser engine, giving Safari preferential access to features compared to competing WebKit-based browsers, limitations placed on in-app browsing, and having Safari pre-installed and prominently displayed as the default browser on iPhones. While users can change the default iPhone web browsing app, investigators say that Safaris designation as the pre-installed default on iPhones reduces user awareness of alternative apps.Investigators found similar concerns regarding Chrome being pre-installed as the default web browser on the vast majority of Android devices. However, the report notes that both Apple and Google have taken steps to make it easier for users to switch to alternative browsers since the investigation announced its provisional findings in November, which have addressed some, but not all, of the concerns relating to choice architecture.The investigation also found that revenue sharing arrangements that see Google paying Apple a significant share of search revenue in exchange for being the default search engine on iPhones was significantly reducing their financial incentives to compete.Apple and Google have yet to respond to our request for comment on the CMAs report.The CMA has put forward potential remedies aimed at improving competition within the UKs mobile browser market, which include forcing Apple to allow developers to use alternative browser engines on iOS, requiring Apple and Google to offer a browser choice screen during device setup, and prohibiting the Chrome revenue sharing arrangements between the two companies. These suggestions are currently unenforceable, however, that could change in the coming months.In January, the CMA launched separate investigations into Apple and Googles mobile ecosystems to decide whether to designate them as having strategic market status (SMS) under the Digital Markets, Competition and Consumers Act (DMCC). Much like the Digital Markets Act (DMA) law introduced by the European Union, the DMCC allows UK regulators to make select companies with substantial and entrenched market power meet stricter antitrust requirements.SMS companies can be imposed with conduct requirements intended to address anticompetitive behavior, and risk fines of up to 10 percent of their annual turnover for violating DMCC rules. If Apple or Google are designated with SMS, the mobile browser investigation is encouraging the CMA to consider imposing appropriate interventions, similar to the suggestions it outlined. The SMS investigations into Google and Apple are currently ongoing and expected to conclude later this year.Following our in-depth investigation, we have concluded that competition between different mobile browsers is not working well, and this is holding back innovation in the UK, said Margot Daly, Chair of the CMAs independent inquiry group. I welcome the CMAs prompt action to open strategic market status investigations into both Apple and Googles mobile ecosystems. The extensive analysis weve set out today will help that work as it progresses.See More:

In this tutorial, we implement a Bilingual Chat Assistant powered by Arcees Meraj-Mini model, which is deployed seamlessly on Google Colab using T4 GPU. This tutorial showcases the capabilities of open-source language models while providing a practical, hands-on experience in deploying state-of-the-art AI solutions within the constraints of free cloud resources. Well utilise a powerful stack of tools including:Arcees Meraj-Mini modelTransformers library for model loading and tokenizationAccelerate and bitsandbytes for efficient quantizationPyTorch for deep learning computationsGradio for creating an interactive web interfaceCopy CodeCopiedUse a different Browser# Enable GPU acceleration!nvidia-smi --query-gpu=name,memory.total --format=csv# Install dependencies!pip install -qU transformers accelerate bitsandbytes!pip install -q gradioFirst we enable GPU acceleration by querying the GPUs name and total memory using the nvidia-smi command. It then installs and updates key Python librariessuch as transformers, accelerate, bitsandbytes, and gradioto support machine learning tasks and deploy interactive applications.Copy CodeCopiedUse a different Browserimport torchfrom transformers import AutoTokenizer, AutoModelForCausalLM, pipeline, BitsAndBytesConfigquant_config = BitsAndBytesConfig( load_in_4bit=True, bnb_4bit_quant_type="nf4", bnb_4bit_compute_dtype=torch.float16, bnb_4bit_use_double_quant=True)model = AutoModelForCausalLM.from_pretrained( "arcee-ai/Meraj-Mini", quantization_config=quant_config, device_map="auto")tokenizer = AutoTokenizer.from_pretrained("arcee-ai/Meraj-Mini")Then we configures 4-bit quantization settings using BitsAndBytesConfig for efficient model loading, then loads the arcee-ai/Meraj-Mini causal language model along with its tokenizer from Hugging Face, automatically mapping devices for optimal performance.Copy CodeCopiedUse a different Browserchat_pipeline = pipeline( "text-generation", model=model, tokenizer=tokenizer, max_new_tokens=512, temperature=0.7, top_p=0.9, repetition_penalty=1.1, do_sample=True)Here we create a text generation pipeline tailored for chat interactions using Hugging Faces pipeline function. It configures maximum new tokens, temperature, top_p, and repetition penalty to balance diversity and coherence during text generation.Copy CodeCopiedUse a different Browserdef format_chat(messages): prompt = "" for msg in messages: prompt += f"<|im_start|>{msg['role']}n{msg['content']}<|im_end|>n" prompt += "<|im_start|>assistantn" return promptdef generate_response(user_input, history=[]): history.append({"role": "user", "content": user_input}) formatted_prompt = format_chat(history) output = chat_pipeline(formatted_prompt)[0]['generated_text'] assistant_response = output.split("<|im_start|>assistantn")[-1].split("<|im_end|>")[0] history.append({"role": "assistant", "content": assistant_response}) return assistant_response, historyWe define two functions to facilitate a conversational interface. The first function formats a chat history into a structured prompt with custom delimiters, while the second appends a new user message, generates a response using the text-generation pipeline, and updates the conversation history accordingly.Copy CodeCopiedUse a different Browserimport gradio as grwith gr.Blocks() as demo: chatbot = gr.Chatbot() msg = gr.Textbox(label="Message") clear = gr.Button("Clear History") def respond(message, chat_history): response, _ = generate_response(message, chat_history.copy()) return response, chat_history + [(message, response)] msg.submit(respond, [msg, chatbot], [msg, chatbot]) clear.click(lambda: None, None, chatbot, queue=False)demo.launch(share=True)Finally, we build a web-based chatbot interface using Gradio. It creates UI elements for chat history, message input, and a clear history button, and defines a response function that integrates with the text-generation pipeline to update the conversation. Finally, the demo is launched with sharing enabled for public access.Here is the Colab Notebook. Also,dont forget to follow us onTwitterand join ourTelegram ChannelandLinkedIn Group. Dont Forget to join our80k+ ML SubReddit.The post Building an Interactive Bilingual (Arabic and English) Chat Interface with Open Source Meraj-Mini by Arcee AI: Leveraging GPU Acceleration, PyTorch, Transformers, Accelerate, BitsAndBytes, and Gradio appeared first on MarkTechPost.

Large language models (LLMs) models primarily depend on their internal knowledge, which can be inadequate when handling real-time or knowledge-intensive questions. This limitation often leads to inaccurate responses or hallucinations, making it essential to enhance LLMs with external search capabilities. By leveraging reinforcement learning, researchers are actively working on methods to improve these models ability to retrieve and integrate relevant information beyond their static knowledge base.Current LLMs restricted access to up-to-date and domain-specific knowledge is a major issue. Since these models are trained on vast datasets that may not include recent developments, they struggle with answering dynamic questions requiring real-time information. While retrieval-augmented generation (RAG) methods have been introduced to mitigate this issue, existing solutions rely heavily on structured prompting and supervised fine-tuning (SFT). These approaches often lead to overfitting, limiting the models generalization ability across different datasets. There is a need for an alternative that allows LLMs to autonomously interact with external search systems, improving their adaptability and accuracy.Previous methods have attempted to incorporate external search functionality into LLMs using iterative prompting, supervised fine-tuning, and tree-based search techniques like Monte Carlo Tree Search (MCTS). While these methods show some improvements, they rely on expensive computational resources and proprietary models. Supervised fine-tuning, for instance, forces models to memorize reasoning paths, which negatively impacts their ability to generalize to new scenarios. Some retrieval-based strategies introduce multi-step query refinement techniques but often require human intervention or predefined prompt templates. These limitations necessitate the development of a more autonomous and efficient search mechanism for LLMs.A research team from the Renmin University of China and DataCanvas Alaya NeW introduced R1-Searcher, a novel reinforcement learning framework designed to improve LLMs ability to retrieve external knowledge effectively. This framework employs a two-stage reinforcement learning approach to enable LLMs to invoke an external search system without requiring human-crafted prompts or prior supervised fine-tuning. By focusing solely on reinforcement learning, R1-Searcher allows models to explore and learn optimal retrieval strategies autonomously, improving accuracy and efficiency in reasoning tasks.The R1-Searcher framework is structured in two phases. The first phase encourages the model to initiate external search actions, providing retrieval-based rewards without considering the final answers correctness. This phase ensures that the model learns to invoke search queries correctly. The second phase refines this capability by introducing an answer-based reward system, which evaluates whether the retrieved information contributes to solving the given problem. The reinforcement learning process relies on a tailored loss function that penalizes incorrect or unnecessary searches while rewarding the effective use of external knowledge. Unlike previous retrieval-based techniques, this approach allows LLMs to integrate reasoning and retrieval dynamically, improving their adaptability across diverse tasks.Experimental evaluations demonstrated that R1-Searcher outperformed existing retrieval-augmented methods, including GPT-4o-mini-based models. On the HotpotQA dataset, accuracy improved by 48.22%, while on the 2WikiMultiHopQA dataset, it achieved a 21.72% increase. Further, it showed strong generalization capabilities by outperforming other models on the Bamboogle dataset, achieving an 11.4% improvement over comparable retrieval-based approaches. Unlike previous techniques, which relied on closed-source models and extensive computational resources, R1-Searcher provided superior performance while maintaining efficiency in search and reasoning tasks. The study also demonstrated that this approach successfully mitigated common issues related to hallucinations and misinformation in LLM-generated responses.The findings indicate that enhancing LLMs with autonomous search capabilities can significantly improve their accuracy and generalization. Using reinforcement learning instead of supervised fine-tuning, R1-Searcher allows models to learn optimal retrieval strategies dynamically, eliminating reliance on memorized responses. This approach represents a major advancement in artificial intelligence, addressing the limitations of existing models while ensuring they remain adaptable to evolving knowledge requirements. The studys results highlight the potential for reinforcement learning to revolutionize knowledge integration in LLMs, making them more reliable for diverse reasoning tasks.Check outthe Paper and GitHub Page.All credit for this research goes to the researchers of this project. Also,feel free to follow us onTwitterand dont forget to join our80k+ ML SubReddit. NikhilNikhil 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.Nikhilhttps://www.marktechpost.com/author/nikhil0980/This AI Paper Introduces RL-Enhanced QWEN 2.5-32B: A Reinforcement Learning Framework for Structured LLM Reasoning and Tool ManipulationNikhilhttps://www.marktechpost.com/author/nikhil0980/Visual Studio Code Setup GuideNikhilhttps://www.marktechpost.com/author/nikhil0980/This AI Paper Introduces CODI: A Self-Distillation Framework for Efficient and Scalable Chain-of-Thought Reasoning in LLMsNikhilhttps://www.marktechpost.com/author/nikhil0980/Getting Started with Kaggle Kernels for Machine Learning Parlant: Build Reliable AI Customer Facing Agents with LLMs (Promoted)