YouZum

Uncategorized

AI, Committee, 新闻, Uncategorized

A woman’s uterus has been kept alive outside the body for the first time

“Think of this as a human body,” says Javier González. In front of me is essentially a metal box on wheels. Standing at around a meter in height, it reminds me of a stainless-steel counter in a restaurant kitchen. It is covered in flexible plastic tubing—which act as veins and arteries—connecting a series of transparent containers, the organs of this machine. What makes it extra special is the role of the cream-colored tub that sits on its surface. Ten months ago, González, a biomedical scientist who developed the device with his colleagues at the Carlos Simon Foundation, carefully placed a freshly donated human uterus in the tub. The team connected it to the device’s tubes and pumped in modified human blood. The device kept the uterus alive for a day—a new feat that could represent the first step to the long-term maintenance of uteruses outside the human body. The work has not yet been published.  The team members want to keep donated human uteruses alive long enough to see a full menstrual cycle. They hope this will help them study diseases of the uterus and learn more about how embryos burrow their way into the organ’s lining at the start of a pregnancy. They also hope that future iterations of their device might one day sustain the full gestation of a human fetus. The machine is technically called PUPER, which stands for “preservation of the uterus in perfusion.” But González’s colleague Xavier Santamaria says the team has adopted a nickname for it: “We call it ‘Mother.’” The organ in the machine González and Santamaria, medical vice president of the Carlos Simon Foundation, demonstrated how the device might work when I visited the foundation in Valencia, Spain, earlier this month (although it held no organs on that day).  Both are interested in learning more about implantation, the moment at which an embryo attaches itself to the lining of a uterus—essentially, the very first moment of pregnancy. The foundation’s founder and director, Carlos Simon, believes it’s a sticking point in IVF: Scientists have made many improvements to the technology over the years, but the failure of embryos to implant underlies plenty of unsuccessful IVF cycles, he says. Being able to carefully study how the process works in a real, living organ might give the team a better idea of how to prevent those failures. JESS HAMZELOU JAVIER GONZALES/CARLOS SIMON FOUNDATION Javier González demonstrates the perfusion machine. A previous iteration of the device kept a sheep’s uterus (right) alive for a day. The team took inspiration from advances in technologies designed to maintain donated organs for transplantation. In recent years, researchers around the world have created devices that deliver nutrients and filter waste so that organs can survive longer after being removed from donors’ bodies. The main goal here is to buy time. A human organ might last only a matter of hours outside the body, so a transplant may require frantic preparation for the recipient, sometimes in the middle of the night. With a little more time, doctors could find better donor-patient matches and potentially test the quality of donated organs. This approach is called normothermic or machine perfusion, and it is already being used clinically for some liver, kidney, and heart transplants. The team at the Carlos Simon Foundation built a similar machine for uteruses. A blood bag hangs on one side. From there, blood is ferried via plastic tubing to a pump, which functions as the heart. The pump shunts the blood through an oxygenator, which adds oxygen and removes carbon dioxide as the lungs would in a human body. The blood is warmed and passed through sensors that monitor the levels of glucose and oxygen, along with other factors. It passes through a “kidney” to remove waste. And finally the blood reaches the uterus, hooked up to its own plastic “arteries” and “veins.” The organ itself sits at a tilt, just as in the body, and is kept in a humid environment to stay moist. Mother’s first uterus The team first began testing an early prototype of the device with sheep uteruses around four years ago. That meant carting the machine to an animal research center in Zaragoza, around 200 miles away. Over the course of the preliminary study, veterinary surgeons removed the uteruses of six sheep and hooked them up to the machine. They kept each uterus alive for a day, using blood from the same animals. After the sheep experiments, the researchers carted their machine back to Valencia and modified it to achieve its current incarnation, “Mother.” They started working with a local hospital that performed hysterectomies. And in May last year, they were offered their first human uterus. The team needed to be quick. “You need to put [the uterus in the machine] within a couple of hours, maximum, of the extraction,” says Santamaria. He and his colleagues also needed to connect the uterus’s blood vessels to the tubing delicately, taking care to avoid any blockages (clotting is a major challenge in organ perfusion). The organ was hooked up to human blood obtained from a blood bank. It seemed to work—at least temporarily. “We kept it alive for one day,” says Santamaria. “As a proof of concept, it is impressive,” says Keren Ladin, a bioethicist who has focused on organ transplantation and perfusion at Tufts University. “These are early days.” It might not sound like much, but 24 hours is a long time for an organ to be out of the body. Maintaining a donated uterus for that long could expand the options for uterus transplant, a fairly new procedure offered to some people who want to be pregnant but don’t have a functional uterus, says Gerald Brandacher, professor of experimental and translational transplant surgery at the Medical University of Innsbruck in Austria. “It is better than what we currently have, because we have only a couple of hours,” he says. So far, most uterus transplants have been planned operations involving organs from living donors.

A woman’s uterus has been kept alive outside the body for the first time Read Post »

AI, Committee, 新闻, Uncategorized

OMIND: Framework for Knowledge Grounded Finetuning and Multi-Turn Dialogue Benchmark for Mental Health LLMs

arXiv:2603.25105v1 Announce Type: new Abstract: Large Language Models (LLMs) have shown remarkable capabilities for complex tasks, yet adaptation in medical domain, specifically mental health, poses specific challenges. Mental health is a rising concern globally with LLMs having large potential to help address the same. We highlight three primary challenges for LLMs in mental health – lack of high quality interpretable and knowledge grounded training data; training paradigms restricted to core capabilities, and evaluation of multi turn dialogue settings. Addressing it, we present oMind framework which includes training and aligning LLM agents for diverse capabilities including conversations; high quality ~164k multi-task SFT dataset, as a result of our generation pipeline based on Structured Knowledge retrieval, LLM based pruning, and review actions. We also introduce oMind-Chat – a novel multi turn benchmark dataset with expert annotated turn level and conversation level rubrics. Our diverse experiments on both core capabilities and conversations shows oMind LLMs consistently outperform baselines. oMind-LLM also shows significantly better reasoning with up to 80% win rate.

OMIND: Framework for Knowledge Grounded Finetuning and Multi-Turn Dialogue Benchmark for Mental Health LLMs Read Post »

AI, Committee, 新闻, Uncategorized

Meta Releases TRIBE v2: A Brain Encoding Model That Predicts fMRI Responses Across Video, Audio, and Text Stimuli

Neuroscience has long been a field of divide and conquer. Researchers typically map specific cognitive functions to isolated brain regions—like motion to area V5 or faces to the fusiform gyrus—using models tailored to narrow experimental paradigms. While this has provided deep insights, the resulting landscape is fragmented, lacking a unified framework to explain how the human brain integrates multisensory information. Meta’s FAIR team has introduced TRIBE v2, a tri-modal foundation model designed to bridge this gap. By aligning the latent representations of state-of-the-art AI architectures with human brain activity, TRIBE v2 predicts high-resolution fMRI responses across diverse naturalistic and experimental conditions. https://ai.meta.com/research/publications/a-foundation-model-of-vision-audition-and-language-for-in-silico-neuroscience/ The Architecture: Multi-modal Integration TRIBE v2 does not learn to ‘see’ or ‘hear’ from scratch. Instead, it leverages the representational alignment between deep neural networks and the primate brain. The architecture consists of three frozen foundation models serving as feature extractors, a temporal transformer, and a subject-specific prediction block. 1. Feature Extraction The model processes stimuli through three specialized encoders: Text: Contextualized embeddings are extracted from LLaMA 3.2-3B. For every word, the model prepends the preceding 1,024 words to provide temporal context, which is then mapped to a 2 Hz grid. Video: The model uses V-JEPA2-Giant to process 64-frame segments spanning the preceding 4 seconds for each time-bin. Audio: Sound is processed through Wav2Vec-BERT 2.0, with representations resampled to 2 Hz to match the stimulus frequency (fstim) (f_{stim}). 2. Temporal Aggregation The resulting embeddings are compressed into a shared dimension (D=384)(D=384) and concatenated to form a multi-modal time series with a model dimension of Dmodel=3×384=1152D_{model} = 3 times 384 = 1152. This sequence is fed into a Transformer encoder (8 layers, 8 attention heads) that exchanges information across a 100-second window. 3. Subject-Specific Prediction To predict brain activity, the Transformer outputs are decimated to the 1 Hz fMRI frequency (ffMRI)(f_{fMRI}) and passed through a Subject Block. This block projects the latent representations to 20,484 cortical vertices (fsaverage5surface)(fsaverage5 surface) and 8,802 subcortical voxels. Data and Scaling Laws A significant hurdle in brain encoding is data scarcity. TRIBE v2 addresses this by utilizing ‘deep’ datasets for training—where a few subjects are recorded for many hours—and ‘wide’ datasets for evaluation. Training: The model was trained on 451.6 hours of fMRI data from 25 subjects across four naturalistic studies (movies, podcasts, and silent videos). Evaluation: It was evaluated across a broader collection totaling 1,117.7 hours from 720 subjects. The research team observed a log-linear increase in encoding accuracy as the training data volume increased, with no evidence of a plateau. This suggests that as neuroimaging repositories expand, the predictive power of models like TRIBE v2 will continue to scale. Results: Beating the Baselines TRIBE v2 significantly outperforms traditional Finite Impulse Response (FIR) models, the long-standing gold standard for voxel-wise encoding. Zero-Shot and Group Performance One of the model’s most striking capabilities is zero-shot generalization to new subjects. Using an ‘unseen subject’ layer, TRIBE v2 can predict the group-averaged response of a new cohort more accurately than the actual recording of many individual subjects within that cohort. In the high-resolution Human Connectome Project (HCP) 7T dataset, TRIBE v2 achieved a group correlation (Rgroup) (R_{group}) near 0.4, a two-fold improvement over the median subject’s group-predictivity. Fine-Tuning When given a small amount of data (at most one hour) for a new participant, fine-tuning TRIBE v2 for just one epoch leads to a two- to four-fold improvement over linear models trained from scratch. In-Silico Experimentation The research team argue that TRIBE v2 could be useful for piloting or pre-screening neuroimaging studies. By running virtual experiments on the Individual Brain Charting (IBC) dataset, the model recovered classic functional landmarks: Vision: It accurately localized the fusiform face area (FFA) and parahippocampal place area (PPA). Language: It successfully recovered the temporo-parietal junction (TPJ) for emotional processing and Broca’s area for syntax. Furthermore, applying Independent Component Analysis (ICA) to the model’s final layer revealed that TRIBE v2 naturally learns five well-known functional networks: primary auditory, language, motion, default mode, and visual. https://aidemos.atmeta.com/tribev2/ Key Takeaway A Powerhouse Tri-modal Architecture: TRIBE v2 is a foundation model that integrates video, audio, and language by leveraging state-of-the-art encoders like LLaMA 3.2 for text, V-JEPA2 for video, and Wav2Vec-BERT for audio. Log-Linear Scaling Laws: Much like the Large Language Models we use every day, TRIBE v2 follows a log-linear scaling law; its ability to accurately predict brain activity increases steadily as it is fed more fMRI data, with no performance plateau currently in sight. Superior Zero-Shot Generalization: The model can predict the brain responses of unseen subjects in new experimental conditions without any additional training. Remarkably, its zero-shot predictions are often more accurate at estimating group-averaged brain responses than the recordings of individual human subjects themselves. The Dawn of In-Silico Neuroscience: TRIBE v2 enables ‘in-silico’ experimentation, allowing researchers to run virtual neuroscientific tests on a computer. It successfully replicated decades of empirical research by identifying specialized areas like the fusiform face area (FFA) and Broca’s area purely through digital simulation. Emergent Biological Interpretability: Even though it’s a deep learning ‘black box,’ the model’s internal representations naturally organized themselves into five well-known functional networks: primary auditory, language, motion, default mode, and visual. Check out the Code, Weights and Demo. Also, feel free to follow us on Twitter and don’t forget to join our 120k+ ML SubReddit and Subscribe to our Newsletter. Wait! are you on telegram? now you can join us on telegram as well. The post Meta Releases TRIBE v2: A Brain Encoding Model That Predicts fMRI Responses Across Video, Audio, and Text Stimuli appeared first on MarkTechPost.

Meta Releases TRIBE v2: A Brain Encoding Model That Predicts fMRI Responses Across Video, Audio, and Text Stimuli Read Post »

AI, Committee, 新闻, Uncategorized

Here’s why some people choose cryonics to store their bodies and brains after death

This week I reported on some rather unusual research that focuses on the brain of L. Stephen Coles. Coles was a gerontologist who died from pancreatic cancer in 2014. He had spent the latter part of his career specializing in human longevity. And before he died, he decided to have his brain preserved by a cryonics facility. Today, it’s being stored at −146 °C at a center in Arizona, where it sits covered in a thin layer of frost. Coles also tasked his longtime friend Greg Fahy with studying pieces of his brain to see how they had fared (partly because he was worried his brain might crack). Fahy, a renowned cryobiologist, has found that the brain is “astonishingly well preserved.” But that doesn’t mean Coles could be reanimated. Over the past few years, I’ve spoken to people who run cryonics facilities, study cryopreservation, or just want to be cryogenically stored. All those I’ve spoken to acknowledge that the chance they’ll one day be brought back to life is vanishingly small. So why do they do it? The first person to be cryonically preserved was James Hiram Bedford, a retired psychology professor who died of kidney cancer in 1967. Affiliates of the Cryonics Society of California, an organization headed by a charming TV repairman with no scientific or medical training, perfused his body with cryoproctective chemicals to protect against harmful ice formation and “quick-froze” him. Today, Bedford’s body is still in storage at Alcor, a cryonics facility based in Scottsdale, Arizona. It’s one of a handful of organizations that offer to collect, preserve, and store a person’s whole body or just their brain—pretty much indefinitely. It’s where Coles’s brain is stored. Both men died from cancer. Medicine could not cure them. But in the future, who knows? One of the premises of cryonics is that modern medicine will continue to advance over time. Cancer death rates have declined significantly in the US since the early 1990s. I don’t know what exactly drove Coles and Bedford to their decisions, but they might have hoped to be reanimated at some point in the future when their cancers became curable. Others simply don’t want to die, period. Last year, I attended Vitalist Bay, a gathering for people who believe that life is good and that death is “humanity’s core problem.” Emil Kendziorra, CEO of the cryonics organization Tomorrow.Bio, spoke at the event, and a healthy interest in cryonics was obvious among the attendees. Many of them believe that science will find a way to “obviate” aging. And some were keen on the idea of being preserved until that happens. Think of it as a way to cheat not only death but aging itself. This sentiment might have support beyond the realms of Vitalist Bay, according to research by Kendziorra and his colleagues. In 2021, they surveyed 1,478 US-based internet users who were recruited via Craigslist. They found that men were more aware of cryonics than women, and more optimistic about its outcomes. Just over a third of the men who completed the survey expressed interest “a desire to live indefinitely.” Still, cryonics is a niche field. Worldwide, only around 5,000 or 6,000 people have signed up for cryopreservation when they die, Kendziorra told me when we chatted at Vitalist Bay. He also told me that his company gets between 20 and 50 new signups every month. And there are plenty of reasons why people don’t do it. A small fraction of the people who responded to Kendziorra’s survey said that they thought the idea of cryonics was dystopian, and some even said it should be illegal. Then there’s the cost. Alcor charges $80,000 to store a person’s brain, and around $220,000 to store a whole body. Tomorrow.Bio’s charges are slightly higher. Many people, including Kendziorra himself, opt to cover this cost via a life insurance policy. Perhaps the main reason people don’t opt for cryonic preservation is that we don’t have any way to bring people back. Bedford has been in storage for more than 50 years, Coles for more than a decade. All the scientists I’ve spoken to say the likelihood of reanimating remains like theirs is vanishingly small. The fact that the possibility—however tiny—is above zero is enough for some, including Nick Llewellyn, the director of research and development at Alcor. As a scientist, he says, he acknowledges that the chances reanimation will actually work are “pretty low.” Still, he’s interested in seeing what the future will look like, so he has signed himself up for the cryonic preservation of his brain. But Shannon Tessier, a cryobiologist at Massachusetts General Hospital, tells me that she wouldn’t sign up for cryonic preservation even if it worked. “It turns into a philosophical question,” she says. “Do I want to be revived hundreds of years later when my family is gone and life is different?” she asks. “There are so many complicated philosophical, societal, [and] legal complications that need to be thought through.” This article first appeared in The Checkup, MIT Technology Review’s weekly biotech newsletter. To receive it in your inbox every Thursday, and read articles like this first, sign up here.

Here’s why some people choose cryonics to store their bodies and brains after death Read Post »

AI, Committee, 新闻, Uncategorized

The Download: the internet’s best weather app, and why people freeze their brains

This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology. How a couple of ski bums built the internet’s best weather app  The best snow-forecasting app for skiers isn’t a federally-funded service or a big-name brand. It’s OpenSnow, a startup that uses government data, its own AI models, and decades of alpine-life experience to deliver the best predictions out there.  The app has proved especially vital this winter, one of the weirdest on record. It’s even made microcelebrities of its forecasters, who sift through reams of data to write “Daily Snow” reports for locations around the world.   We headed to the Tahoe mountains to hear how two broke ski bums became modern-day snow gods. Read the full story.  —Rachel Levin  Here’s why some people choose cryonics to store their bodies and brains after death  —Jessica Hamzelou  This week I reported on unusual research focused on the frozen brain of L. Stephen Coles.  Coles, a researcher who studied aging, was interested in cryonics—the long-term storage of human bodies and brains in the hope that they might one day be brought back to life. It’s a hope shared by many.  Over the past few years, I’ve spoken to people who run cryonics facilities, study cryopreservation, or just want to be cryogenically stored. All of them acknowledge that there’s a vanishingly small chance of being brought back to life. So why do they do it?  Read the full story to find out.  This article is from The Checkup, our weekly biotech newsletter. Sign up to receive it in your inbox every Thursday.  What’s next for space exploration?   Whether it’s the race to find life on Mars, the campaign to outsmart killer asteroids, or the quest to make the moon a permanent home to astronauts, scientists’ efforts in space can tell us more about where humanity is headed. To learn more about the progress and possibilities ahead, our features editor Amanda Silverman sat down with Robin George Andrews, an award-winning science journalist and author, on Wednesday. If you missed their conversation, fear not—you can catch up and watch the video here. You’ll need to be a subscriber to access it, but the good news is subscriptions are discounted right now. Bag yours if you haven’t already!  The must-reads  I’ve combed the internet to find you today’s most fun/important/scary/fascinating stories about technology.  1 The Pentagon’s ban on Anthropic has been halted A judge has paused its designation as a supply chain risk. (CBS News)  + She said the government was trying to “chill public debate.” (BBC) + Sam Altman claimed he tried to “save” Anthropic in the clash. (Axios)  2 Elon Musk has lost his lawsuit against an ad boycott on X A judge admonished the “fishing expedition.” (Ars Technica) + Ad revenue fell by more than half as advertisers fled X after Musk took over. (BBC)  3 OpenAI has put plans for an erotic chatbot on hold “indefinitely” Staff and investors had raised concerns. (The Information $) + The company is making a sharp strategic pivot. (FT $) + AI companions are the final stage of digital addiction. (MIT Technology Review)  4 A helium shortage has started impacting tech supply chains The problem stems from the Middle East conflict. (Reuters) + The era of cheap helium is over. (MIT Technology Review)  5 Trump’s new science advisers: 12 tech chiefs and just one academic They include at least nine billionaires. (Nature) + David Sacks is stepping down as Trump’s crypto and AI czar. (TechCrunch)  6 Anthropic is mulling an IPO as soon as October It’s racing OpenAI to hold an initial public offering. (Bloomberg $)  7 Wikipedia has banned all AI-generated content  LLM-related issues had overwhelmed editors. (404 Media) + Here’s what we’re getting wrong about AI’s truth crisis. (MIT Technology Review)  8 OpenAI’s ad pilot generated $100 million in under 2 months More than 600 advertisers are working on the trial. (CNBC) + Ads will arrive on ChatGPT free ‌and Go in the coming weeks. (Reuters)   9 An Irish village is giving kids a phone-free upbringing The ban works because almost everyone’s bought in. (NYT $)  10 Chatting with sycophantic AI makes you less kind New research found it encourages “uncouth behavior.” (Nature)  Quote of the day  “I don’t know if it’s ‘murder,’ but it looks like an attempt to cripple Anthropic.”  —Judge Rita Lin rules against the Pentagon’s ban on Anthropic, The Verge reports.  One More Thing  AURELIA INSTITUTE This futuristic space habitat is designed to self-assemble in orbit   More and more people are traveling beyond Earth, but the International Space Station can only hold 11 of them at a time.   Aurelia Institute, an architecture R&D lab based in Cambridge, MA, is building a solution: a habitat that launches in compact stacks of flat tiles—and self-assembles in orbit.   The concept may sound far-fetched, but it’s already won support from NASA. Read the full story.  —Sarah Ward  We can still have nice things  A place for comfort, fun and distraction to brighten up your day. (Got any ideas? Drop me a line.)  + These optical illusions are absolute brain-melters. + The web design museum lovingly visualizes the evolution of the internet. + Zara Picken’s modernist illustrations are a new window into the mid-20th century. + Explore our planet’s connections through the digital Knowledge Garden. 

The Download: the internet’s best weather app, and why people freeze their brains Read Post »

AI, Committee, 新闻, Uncategorized

Did You Forget What I Asked? Prospective Memory Failures in Large Language Models

arXiv:2603.23530v1 Announce Type: new Abstract: Large language models often fail to satisfy formatting instructions when they must simultaneously perform demanding tasks. We study this behaviour through a prospective memory inspired lens from cognitive psychology, using a controlled paradigm that combines verifiable formatting constraints with benchmark tasks of increasing complexity. Across three model families and over 8,000 prompts, compliance drops by 2-21% under concurrent task load. Vulnerability is highly type-dependent: terminal constraints (requiring action at the response boundary) degrade most, with drops up to 50%, while avoidance constraints remain comparatively robust. A salience-enhanced format (explicit instruction framing plus a trailing reminder) recovers much of the lost compliance, restoring performance to 90-100% in many settings. Interference is bidirectional: formatting constraints can also reduce task accuracy, with one model’s GSM8K accuracy dropping from 93% to 27%. In additional stacking experiments, joint compliance declines sharply as constraints accumulate. All results use deterministic programmatic checkers without an LLM-as-judge component on publicly available datasets.

Did You Forget What I Asked? Prospective Memory Failures in Large Language Models Read Post »

AI, Committee, 新闻, Uncategorized

Tencent AI Open Sources Covo-Audio: A 7B Speech Language Model and Inference Pipeline for Real-Time Audio Conversations and Reasoning

Tencent AI Lab has released Covo-Audio, a 7B-parameter end-to-end Large Audio Language Model (LALM). The model is designed to unify speech processing and language intelligence by directly processing continuous audio inputs and generating audio outputs within a single architecture. System Architecture The Covo-Audio framework consists of four primary components designed for seamless cross-modal interaction: Audio Encoder: The model utilizes Whisper-large-v3 as its primary encoder due to its robustness against background noise and varied accents. This component operates at a frame rate of 50 Hz. Audio Adapter: To bridge the encoder and the LLM, a specialized adapter employs three downsampling modules, integrating linear and convolution layers to reduce the frame rate from 50 Hz to 6.25 Hz. LLM Backbone: The system is built upon Qwen2.5-7B-Base, which has been adapted to process interleaved sequences of continuous acoustic features and textual tokens. Speech Tokenizer and Decoder: The tokenizer, based on WavLM-large, uses a codebook size of 16,384 to produce discrete audio tokens at 25 Hz. The decoder employs a Flow-Matching (FM) based framework and a BigVGAN vocoder to reconstruct high-fidelity 24K waveforms. https://arxiv.org/pdf/2602.09823 Hierarchical Tri-modal Interleaving A core contribution of this work is the Hierarchical Tri-modal Speech-Text Interleaving strategy. Unlike traditional methods that operate solely at the word or character level, this framework aligns continuous acoustic features (ac)(a_c), discrete speech tokens (ad)(a_d), and natural language text (t)(t). The model utilizes two primary patterns: Sequential Interleaving (ac→t→ad)(a_c rightarrow t rightarrow a_d): Continuous features, text, and discrete tokens are arranged in a progressive chain. Parallel Integration (ac→t|ad)(a_c rightarrow t | a_d): Continuous features are aligned with a coupled text-discrete unit. The hierarchical aspect ensures structural coherence by using phrase-level interleaving for fine-grained alignment and sentence-level interleaving to preserve global semantic integrity in long-form utterances. The training process involved a two-stage pre-training pipeline processing a total of 2T tokens. Intelligence-Speaker Decoupling To mitigate the high cost of constructing large-scale dialogue data for specific speakers, the research team proposed an Intelligence Speaker Decoupling strategy. This technique separates dialogue intelligence from voice rendering, allowing for flexible voice customization using minimal text-to-speech (TTS) data. The method reformats high-quality TTS recordings into pseudo-conversations with masked text loss. By excluding the text response portion from the loss calculation, the model preserves its reasoning abilities while inheriting the naturalness of the TTS speaker. This enables personalized interaction without the need for extensive, speaker-specific dialogue datasets. Full-Duplex Voice Interaction Covo-Audio evolved into Covo-Audio-Chat-FD, a variant capable of simultaneous dual-stream communication. The audio encoder is reformatted into a chunk-streaming manner, and the user and model streams are chunk-interleaved in a 1:4 ratio. Each chunk represents 0.16s of audio. The system manages conversational states through specific architectural tokens: THINK Token: Indicates a listening-only state while the model waits to respond. SHIFT Token: Signifies the transition to the model’s speaking turn. BREAK Token: Detects interruption signals (barge-ins), triggering the model to terminate speaking immediately and switch back to listening. For multi-turn scenarios, the model implements a recursive context-filling strategy, where continuous audio features from user input and generated tokens from previous turns are prefixed as historical context. Audio Reasoning and Reinforcement Learning To enhance complex reasoning, the model incorporates Chain-of-Thought (CoT) reasoning and Group Relative Policy Optimization (GRPO). The model is optimized using a verifiable composite reward function: $$R_{total} = R_{accuracy} + R_{format} + R_{consistency} + R_{thinking}$$ This structure allows the model to optimize for correctness (Raccuracy)(R_{accuracy}), structured output adherence (Rformat)(R_{format}), logical coherence (Rconsistency)(R_{consistency}), and reasoning depth (Rthinking)(R_{thinking}). Evaluation and Performance Covo-Audio (7B) shows competitive or superior results on several evaluated benchmarks, with strongest claims made for models of comparable scale and selected speech/audio tasks. On the MMAU benchmark, it achieved an average score of 75.30%, the highest among evaluated 7B-scale models. It notably excelled in music understanding with a score of 76.05%. On the MMSU benchmark, Covo-Audio achieved a leading 66.64% average accuracy. Regarding its conversational variants, Covo-Audio-Chat demonstrated strong performance on URO-Bench, particularly in speech reasoning and spoken dialogue tasks, outperforming models like Qwen3-Omni on the Chinese track. For empathetic interaction on the VStyle benchmark, it achieved state-of-the-art results in Mandarin for anger (4.89), sadness (4.93), and anxiety (5.00). The research team notes an ‘early-response’ issue on the GaokaoEval full-duplex setting, where unusually long silent pauses between vocal fragments can cause premature responses. This ‘early-response’ behavior correlates with the model’s pause-handling success metric and is identified as a critical direction for future optimization. Key Takeaways Unified End-to-End Architecture: Covo-Audio is a 7B-parameter model that natively processes continuous audio inputs and generates high-fidelity audio outputs within a single, unified architecture. It eliminates the need for cascaded ASR-LLM-TTS pipelines, reducing error propagation and information loss. Hierarchical Tri-modal Interleaving: The model employs a specialized strategy to align continuous acoustic features, discrete speech tokens, and natural language text. By interleaving these modalities at both phrase and sentence levels, it preserves global semantic integrity while capturing fine-grained prosodic nuances. Intelligence-Speaker Decoupling: Tencent research team introduces a technique to decouple dialogue intelligence from specific voice rendering. This allows for flexible voice customization using lightweight Text-to-Speech (TTS) data, significantly lowering the cost of developing personalized conversational agents. Native Full-Duplex Interaction: The Covo-Audio-Chat-FD variant supports simultaneous listening and speaking. It utilizes specific architectural tokens—THINK, SHIFT, and BREAK—to manage complex real-time dynamics such as smooth turn-taking, backchanneling, and user barge-ins. Superior Parameter Efficiency: Despite its compact 7B scale, Covo-Audio achieves state-of-the-art or highly competitive performance across core benchmarks, including MMAU, MMSU, and URO-Bench. It frequently matches or exceeds the performance of much larger systems, such as 32B-parameter models, in audio and speech understanding tasks. Check out the Paper, Model on HF and Repo. Also, feel free to follow us on Twitter and don’t forget to join our 120k+ ML SubReddit and Subscribe to our Newsletter. Wait! are you on telegram? now you can join us on telegram as well. The post Tencent AI Open Sources Covo-Audio: A 7B Speech Language Model and Inference Pipeline for Real-Time Audio Conversations and Reasoning appeared first on MarkTechPost.

Tencent AI Open Sources Covo-Audio: A 7B Speech Language Model and Inference Pipeline for Real-Time Audio Conversations and Reasoning Read Post »

AI, Committee, 新闻, Uncategorized

The snow gods: How a couple of ski bums built the internet’s best weather app

The best snow-forecasting app for skiers and snowboarders isn’t from any of the federally funded weather services. Nor from any of the big-name brands. It’s an independent app startup that leverages government data, its own AI models, and decades of alpine-life experience to offer better snow (and soon avalanche) predictions than anything else out there. Skiers in the know follow OpenSnow and won’t bother heading to the mountains—from Alpine Meadows to Mont Blanc, Crested Butte to Killington—unless this small team of trusted weathered men tells them to. (And yes, they’re all men.) The app has made microcelebrities of its forecasters, who sift through and analyze reams of data to write “Daily Snow” reports for locations throughout the world. “I’m F-list famous,” OpenSnow founding partner and forecaster Bryan Allegretto says with a laugh. “Not even D-list.”  The app has proved especially vital this year, which has been one of the weirder winters on record. The US West saw very little daily snow, despite an intense storm cycle that led to one of the deadliest avalanches in history. That storm was followed by one of the fastest melts in memory, and several resorts in California are already shutting down for the season. Meanwhile, in the East, the ongoing snowfall has offered a rare gift: a deep and seemingly endless winter..  MIT Technology Review caught up with Allegretto, better known as BA, in the Tahoe mountains to talk about the weather, AI, avalanches, and how a little weather app became the closest thing powder-hounds have to a crystal ball: a daily dump of the freshest, most decipherable, and most micro-accurate forecasts in the biz. And how two once-broke ski bums—Allegretto and his Colorado counterpart, CEO Joel Gratz— managed to bootstrap a business and turn an email list of 37 into a cult following half a million strong.  This interview has been edited for clarity and accuracy.  You grew up in New Jersey. Middle of the pack as far as snowy states. What were your winters like as a kid? I was always obsessed with weather. Especially severe weather. Nor’easters. There was the blizzard of ’89, I believe, that hit the East Coast hard—dropped two to three feet of snow, which was a lot for the Jersey Shore. My dad worked for the highway authority, so he had tools other than the evening news. He was in charge of calling out the snowplows whenever it snowed, so I just remember chasing storms with my dad. I wasn’t allowed to ride in the snowplows. I’d watch them. When I got older, I was the one shoveling the neighbors’ driveways. I just liked being out there. In it. In college, I used to go around and shovel all the girls’ sidewalks. That was fun.  When did you start skiing? We would cut school and take a bus to go skiing, unbeknownst to our parents. It was the ’90s, and the surfers decided snowboarding would be fun, so the local surf shop started  running a bus and all these surfers would show up and hop the bus to Hunter Mountain. We’d drive to the Poconos, go night skiing, turn around. It wasn’t uncommon for me in high school to get in the car by myself, either —and just drive. Me, my dog, my backpack. I’d sleep in gas stations and ski. Storm-chasing around the Northeast.  What were you really chasing, you think? Natural highs. Happiness. I’ve always been a soul-searcher. I grew up in a crazy house situation, a broken home. My dad left. My mom became a drug addict. I just wanted to be gone. I’m the oldest. I was always trying to help my mom and make sure she was okay. No one was telling me to go to school and have a career. I just wanted to do something that fulfills me. How’d you go about figuring out what that was?  For me, to go to school was a big task, given where I was coming out of. There wasn’t any money. I could get grants and scholarships because my mom was so poor. I wanted to go to Penn State but didn’t have the grades. I ended up at Kean, a public university in New Jersey. It had a meteorology program. We got to go to New York City, to NBC, and practiced on the green screen. In meteorology school, I started thinking: How do I work in the ski and snowboard industry and use weather at the same time? I went to Rowan [University] for business, in South Jersey, and in between moved to Hawaii to surf and spent a year teaching snowboarding. My goal the whole time was to not work in a career I hated. I imagine you weren’t like most meteorology students.  Us punk rockers, skaters, snowboarders—we were a little different than the typical meteorology nerds. I was the radical storm chaser. A big personality. I still am. You didn’t quite fit the traditional weatherman mold. Back then, there were no smartphones or social media. If you were a meteorologist, you either worked in a cubicle for the government or at an insurance company assessing weather risk.  Or you were on the local news. That wasn’t my thing. They didn’t want Grizzly Adams up there with his big beard. Beards belong in the mountains? Meteorologists live in cities because that’s where the jobs are. They don’t live in small mountain towns.  That’s what was missing in the industry. When I moved to Tahoe, in 2006, I realized nobody had any trust in the weather forecasts. It was more like a “We’ll believe it when we see it” old-fashioned mentality. If you’re a forecaster in flat areas, you just look at the weather model and regurgitate the news. Weathermen in Sacramento or Reno didn’t give a crap about the ski resorts! They’d just say “We’ll see three feet above 6,000 feet” and go on to the next segment. And skiers were like: “Wait a minute. Is it going to

The snow gods: How a couple of ski bums built the internet’s best weather app Read Post »

AI, Committee, 新闻, Uncategorized

Are high gas prices good news for EVs? It’s complicated.

I live in a dense city with plentiful public transportation options and limited parking, so I don’t own a car. I’m often utterly clueless about the current price of gasoline. But as the conflict in Iran has escalated, fossil-fuel prices have been on a roller-coaster, and I’ve started paying attention. In the US, average gas prices are $3.98 a gallon as of March 25, up from under $3 before the war started. Online there’s been what almost looks like cheerleading about this volatility from some folks, including EV owners—some of the social media posts and op-eds have read as nearly gleeful. The subtext (or even the text) is “I told you so.”  Don’t get me wrong—this could be an opportunity for EVs to make headway around the world. But there are plenty of reasons that even the carless among us should be concerned about a sustained rise in fossil-fuel prices. Historically, this is exactly the sort of moment that’s pushed people to reevaluate how they get around. During the oil crisis of the 1970s, Americans switched to smaller, more efficient cars in droves. It was a major opportunity for Japanese automakers, whose vehicles tended to fit this mold better than those produced by their US counterparts. We’re already seeing early signs that people are interested in going electric. One US-based online car marketplace said that search traffic for EVs was up 20% following the initial attack on Iran. For more popular models like the Tesla Model Y, traffic nearly doubled. And the interest is global. One car dealership outside London said it’s struggling to keep up with demand and is sending staff to buy more EVs at auction, according to Reuters. Another in Manila told Bloomberg that it got a month’s worth of orders in two weeks. The timing here is really interesting in the US in particular, because we’re about to see a wave of more affordable used EVs hit the market. Three years ago, a leasing boom started with the Inflation Reduction Act, which included incentives for EVs, including leases. About 300,000 such leases are set to expire this year, and many of those vehicles could come up for sale, increasing the available supply of affordable used EVs. The interest is there, but what would it really take for more drivers to make the switch? Nice, round numbers do tend to get people’s attention. Some point to $4 per gallon (which the national average is quite close to right now). At that price, the total cost of ownership for an EV is comfortably lower than the cost for a gas-powered car, even with higher electricity prices, according to data from the energy consultancy BloombergNEF. Then again, maybe that won’t quite do the trick: One survey from Cox Automotive found that most US consumers would consider switching to an EV or hybrid if gas prices hit $6 per gallon. But this is also the second big incident of fossil-fuel volatility in the last five years, which could make consumers more ready to make the switch, as Elaine Buckberg, a senior fellow at Harvard, told Bloomberg. (The first was in the summer of 2022 when Russia invaded Ukraine.) I’m a climate and energy reporter, and I care about addressing climate change. So I’m always happy to hear about people shifting to EVs or any other option that helps cut down on greenhouse-gas emissions. But one aspect that I think is getting lost here is that sustained high fossil-fuel prices will be bad for even those of us who are untethered from the burdens of vehicle ownership. Fuel cost makes up between 50% and 60% of the cost of shipping goods overseas. Fertilizer production today requires natural gas, which has gotten significantly more expensive since the war began, particularly in Europe. Jet fuel prices have basically doubled in the last month, according to the International Air Transport Association. Since those prices account for something like a quarter of an airline’s operating cost, that could soon make air travel—and anything that’s shipped by plane—more expensive. And if all this adds up to an economic downturn, it’s bad for big projects that need financing (even wind and solar farms) and for people who want to borrow money to buy a home or a car (including an EV). If you’re in the market for a car, maybe this uncertainty is what you needed to consider electric. But until we’re able to truly decarbonize not only our transportation but the rest of our economy, even this carless reporter is going to be worried about high gas prices. This article is from The Spark, MIT Technology Review’s weekly climate newsletter. To receive it in your inbox every Wednesday, sign up here. 

Are high gas prices good news for EVs? It’s complicated. Read Post »

We use cookies to improve your experience and performance on our website. You can learn more at 隱私權政策 and manage your privacy settings by clicking Settings.

Privacy Preferences

You can choose your cookie settings by turning on/off each type of cookie as you wish, except for essential cookies.

Allow All
Manage Consent Preferences
  • Always Active

Save
zh_CN