Small Language Models (SLMs) are a cost-effective alternative to Large Language Models (LLMs), but often struggle with complex reasoning due to their limited capacity and a tendency to produce mistakes or inconsistent answers during multi-step reasoning. Existing efforts have improved SLM performance, but typically at the cost of one or more of three key aspects: (1) reasoning capability, due to biased supervision that filters out negative reasoning paths and limits learning from errors; (2) autonomy, due to over-reliance on externally generated reasoning signals; and (3) generalization, which suffers when models overfit to teacher-specific patterns. In this paper, we introduce ReaLM, a reinforcement learning framework for robust and self-sufficient reasoning in vertical domains. To enhance reasoning capability, we propose Multi-Route Process Verification (MRPV), which contrasts both positive and negative reasoning paths to extract decisive patterns. To reduce reliance on external guidance and improve autonomy, we introduce Enabling Autonomy via Asymptotic Induction (EAAI), a training strategy that gradually fades external signals. To improve generalization, we apply guided chain-of-thought distillation to encode domain-specific rules and expert knowledge into SLM parameters, making them part of what the model has learned. Extensive experiments on both vertical and general reasoning tasks demonstrate that ReaLM significantly improves SLM performance across aspects (1)-(3) above.

The moon has turned to cheese in John Scalzi's new sci-fi novel My only complaint about the science fiction due to be published in March is: how in the world are we meant to find the time to read all these great novels? There are so many must-reads out this month, whether it's the latest from Nicholas Binge, Silvia Park's tale of a lost robot sibling or Laila Lalami's vision of a future where our dreams are policed for what we might be going to do (sounds quite Minority Report – a very good thing in my view). All I can say is, I think it's time to step away from the computer and get reading, if we want to keep up… Sadly for humanity, in this latest slice of comic sci-fi from the excellent John Scalzi, the moon has turned to cheese and they have to work out what to do about it. This sounds like a lot of fun, but I'm primarily planning to read it to find out what type of cheese the moon has become. Our sci-fi columnist Emily H. Wilson heartily approves of Binge's latest, writing that this time travel tale is well-deserving of its upcoming big-screen treatment.

Human matting is a foundation task in image and video processing, where human foreground pixels are extracted from the input. Prior works either improve the accuracy by additional guidance or improve the temporal consistency of a single instance across frames. We propose a new framework MaGGIe, Masked Guided Gradual Human Instance Matting, which predicts alpha mattes progressively for each human instances while maintaining the computational cost, precision, and consistency. Our method leverages modern architectures, including transformer attention and sparse convolution, to output all instance mattes simultaneously without exploding memory and latency. Although keeping constant inference costs in the multiple-instance scenario, our framework achieves robust and versatile performance on our proposed synthesized benchmarks. With the higher quality image and video matting benchmarks, the novel multi-instance synthesis approach from publicly available sources is introduced to increase the generalization of models in real-world scenarios.

This story is part of Future Tense Fiction, a monthly series of short stories from Future Tense and Arizona State University's Center for Science and the Imagination about how technology and science will change our lives. Amanda sat at her desk, picking at the same $30 Little Gem salad she ordered daily, suffering a small burning sensation in her gut that was triggered either by acid reflux or the dying embers of her rapidly expiring conscience. Of course, it was standard procedure for her husband to demand that the security firm Dark Metal surveil potential new hires for any of his multibillion-dollar companies, but this was the first time Amanda had been involved in contracting the private intelligence agency herself. Seedlings is your venture, Reid had promised her, even though he'd named himself CEO. I want you to take the lead on this. Amanda was COO of Seedlings and reported to her husband, who dismissed Amanda's concerns about the legal ramifications of their actions. Worrying about the law was something poor people did, Reid insisted. Besides, she'd never seen Reid do anything that nefarious with this type of information. But Maggie Everett was the type of candidate that pleased Reid. Amanda had done her job, which was to find Maggie, and the people at Dark Metal had done theirs, which was to surveil her and create a comprehensive biographical profile. This seemed like overkill to Amanda. Maggie wasn't in the running to become a high-profile executive at one of Reid's billion-dollar firms. She was being interviewed to work at a preschool. Certainly, Seedlings differed from other private preschools--there was the possibility Maggie would be exposed to confidential information. But this was what NDAs were for. Unleashing a network of spies upon a poor teacher who would ultimately be responsible for 10 toddlers seemed like an absurd waste of resources. And this was just Phase 1. Phase 2 would have to wait until after Maggie was hired, of course. Amanda reopened Dark Metal's inch-thick dossier. The logline: Maggie was smart but stupid. Smart: She'd majored in English at Yale, then received an MFA in creative writing from Brown, and finally a master's in early childhood education from Columbia. Stupid: She'd accumulated $103,345 in student debt, which she'd never pay off unless she took a job somewhere like Seedlings.

Recently, I was looking for a toy dataset for my new book's chapter (you can subscribe to the updates here) on instance segmentation. And, I really wanted to have something like the Iris Dataset for Instance Segmentation so that I would be able to explain the model without worrying about the dataset too much. But, alas, it is not always possible to get a dataset that you are looking for. I actually ended up looking through various sources on the internet but inadvertently found that I would need to download a huge amount of data to get anything done. Given that is not at all the right way to go about any tutorial, I thought why not create my own dataset.

For many founders the road to raising capital usually takes months, if not years, but for Flamingo founder Dr Catriona Wallace the latest part of her journey took just 12 minutes. The artificial intelligence fintech startup, which is listed on the Australian Securities Exchange through its holding company Cre8tek, raised $5.1 million last month via a share placement to new and existing institutional, professional and sophisticated investors. The oversubscribed round closed after just 12 minutes, with $10 million in bids vying for the 128.1 million in shares that were on offer for 4c each. "There's recently been a lot more interest and conversation [in Australia] around artificial intelligence, and this is reflected in the success of this raise," Wallace tells StartupSmart. "Like all startup founders I know that raising money is always a difficult task -- we were very pleased that the $5.1 million was oversubscribed and mainly our feeling was of happiness and confidence. It gives us a really significant runway into 2018 to prove our next key milestones."

"Supergirl" stars Chyler Leigh and Floriana Lima revealed that they were pleasantly surprised by how significant Alex and Maggie's relationship has become to viewers who are struggling with their sexuality. "There was no way we would have known the impact that this would have had," Leigh told People at the 28th GLAAD Media Awards ceremony, where "Supergirl" was nominated for best dramatic TV show for its story featuring the lesbian relationship of Alex and Maggie. "We definitely wanted [Alex and Maggie's relationship] to be a strong representation, and that's why we've thought so hard about it and wanted it to be beautifully done," Leigh continued. "So I'm just really happy with the writers of'Supergirl,' who have really brought this to life." READ: Is there a James-centric "Supergirl" episode coming up? "We're just really humbled by it, very proud of it," added Lima who attended the awards night with Leigh.

"Maggie is a very smart monkey," says Tim Buschman, a graduate student in Professor Earl Miller's neuroscience lab. Maggie isn't visible – she's in a biosafety enclosure meant to protect her from human germs – but the signs of her intelligence flow over two monitors in front of Buschman. For the last seven years, Maggie has "worked" for MIT's Department of Brain and Cognitive Sciences (BCS). Three hours a day, the macaque plays computer games that (usually) are designed to require her to generate abstract representations and then use those abstractions as tools. "Even I have trouble with this one," Buschman says, nodding at a game that involves classifying logical operations. But Maggie is on a roll, slamming through problems, taking about half a second for each and getting about four out of five right.

Binary is the story of a man who must decide between saving his artificially intelligent girlfriend or a human woman during a crucial space voyage, causing him to question what love really means. Binary will be a 20-minute, live-action short film. Ryker - Loyal and naive, Ryker grew up in the working class. When he falls in love with an artificially intelligent woman, Seline, he struggles with whether to prioritize his love or his fellow human. Seline - An advanced form of artificial intelligence, appearing human, Seline has been programmed with the capacity to love. She must convince Ryker and Natalie her love is as valuable as human love.