AI·Nubia is set to launch the second-generation 'Doubao Phone' at WAIC 2026 as a mass-market flagship with improved AI agent OS, memory, and execution capabilities.
Search
What are you looking for?
25 results for “generation”
AI·The author benches a quad 5060Ti setup for Qwen3.6-27B code generation using MTP, reporting 608 tokens/s prefill and 52.2 tokens/s decode at 256k context on a cost-effective build.
AI·This position paper reviews LLM-driven formal theorem provers and argues that current systems function mainly as solvers for well-defined problems, not as research agents capable of discovering new theorems or resolving open conjectures at the frontier. It identifies key limitations in datasets, exploration, tools, and collaboration and proposes a roadmap for AI4Math systems to support genuine mathematical research.
AI·MASTE: A Multi-Agent Pipeline for Zero-Shot Aspect Sentiment Triplet Extraction
AI·XALPHA is a memory-driven AI quant researcher that uses multi-source research memory integrating external financial reports and prior discovery feedback, with Macro Brain for theme planning and archetype selection, Micro Brain for hypothesis-to-code translation and tri-alignment verification, and Cross Brain for feedback consolidation, enabling closed-loop continuous alpha discovery that outperforms baselines on CSI300.
AI·Researchers developed SQZ-Qwen-ASR-1.7B for MLC-SLM 2026 Task 1, combining a modular speaker diarization front-end (VAD, CAMPPlus embeddings, spectral clustering, RTTM segmentation) with Qwen3-ASR-1.7B adapted via full supervised fine-tuning, LoRA on TTS-generated synthetic speech, and GRPO reinforcement learning for lower tcpMER (23.70 on dev set, 17.97 on eval set).
AI·COBART uses a bidirectional auto-regressive transformer for optimized ad headline generation with multi-objective control over quality and CTR.
Scientific ideas rarely start from a blank page. They inherit mechanisms, repair known limitations, and recombine pieces of earlier work, much like biological genomes. Current benchmarks still say little about whether AI systems can follow this inheritance structure. We present IdeaGene-Bench (IG-Bench), a benchmark for scientific lineage reasoning and lineage-grounded idea generation. IG-Bench is organized around the IdeaGene framework: each paper or proposal is represented as a set of minimal…
In this work, we present Canvas360, a two-stage framework for in-context panoramic generation that combines geometry-aware pretraining with downstream task-specific fine-tuning. To address the lack of large-scale, high-quality training data tailored to in-context panoramic tasks, we propose Canvas360Dataset, a collection of 1M high-quality paired panoramic samples for style transfer, inpainting, outpainting, and editing, enabling effective supervision across diverse in-context generation scenar…
Reasoning has become a core capability for large models, especially when reliable decisions require understanding logical consequences. Recent video generation models offer a reasoning path distinct from previous Chain-of-Thought (CoT): reasoning can unfold through temporally connected frames, known as Chain-of-Frame (CoF) reasoning. However, existing video generators are primarily trained on general video corpora, still lacking diverse supervision and dedicated designs for CoF reasoning. To ad…
Generating realistic 3D human motions in real-time within interactive applications is key for animation, simulation, and humanoid robotics. While recent offline motion generation approaches offer precise control via text and kinematic constraints, they lack the inference speed required for interactive settings. Conversely, existing online methods enable real-time synthesis but often sacrifice controllability or struggle with complex text semantics and long-horizon goals due to limited context w…
Semantic audio applications increasingly require controllable generation on commodity and embedded hardware rather than through framework-heavy datacenter stacks. We present aria, a dependency-free native runtime that runs the complete text-to-music pipeline of Stable Audio~3 (SA3) on ordinary GPUs, CPU-only machines, and a Raspberry~Pi~5, with no Python or deep-learning framework underneath. Our main contribution is a study of quantization: running the model at lower numerical precision to fit…
Modern LLMs are increasingly deployed in long-context applications such as retrieval-augmented generation, repository-level coding, and agentic workflows whose accumulated reasoning and tool traces routinely push the input an order of magnitude past the pretraining window, making zero-shot context extension the dominant deployment path for open-weight checkpoints. Most existing zero-shot methods fix a single rescaling factor up front, so an aggressive factor sacrifices short-context fidelity wh…
A new generation of voice models for natural human-AI interaction, now powering ChatGPT Voice.
Coding agents increasingly generate pull requests (PRs) for real-world software issues, yet one-shot PR generation remains open-loop: the PR is proposed without systematic review, diagnosis, or revision. We introduce SWE-Review, a framework for closing this loop with agentic code review. Given an issue and an AI-generated PR, a reviewer agent explores the repository, decides whether the PR should be accepted, and provides structured feedback for revision. We evaluate this setting with our propo…
OpenAI previews GPT-5.6 Sol, a next-generation model with stronger capabilities in coding, science, and cybersecurity, paired with its most advanced safety stack.
Our newest audio model introduces granular audio tags that give you precise control to direct AI speech for expressive audio generation.
Doing a bachelor thesis on fine-grained car classification (telling apart VW Golf generations from listing photos). Simple setup: frozen encoder → embeddings → weighted k-NN. On my small dataset (175 train / 132 test): I thought maybe it was a cosine vs euclidean thing, but my embeddings are L2-normalized so both give the same ranking. Tried both, DINOv2 stays at 41%. I get that SigLIP was traine…
Inference-time scaling for text-to-image generation has progressed from simple Best-of-N (BoN) sampling to guided search methods that verify and steer candidate trajectories at intermediate denoising steps. These approaches focus on when and how often to verify during denoising but largely treat the cost of generation itself as fixed. Moreover, the standard practice of comparing methods by number of function evaluations (NFEs) counts only denoising forward passes and ignores verifier overhead, …
Geometry-conditioned 3D scene generation enables the creation of 3D environments from user-provided geometry, offering direct control over scene structure and object layout. To generate such 3D scenes, current methods commonly adopt a three-stage design that first defines a view schedule, then synthesizes multi-view observations along the scheduled views, and finally reconstructs a 3D representation from the generated images. However, defining the view schedule becomes a major bottleneck for ou…
Multimodal large language models (MLLMs) generate responses autoregressively, integrating visual and linguistic information in an evolving context. Prior work on interpretability has focused on individual layers and circuits (where), leaving the token-level dynamics of multimodal computation during generation (when) underexplored. We address this gap and study attention shifts as per semantic role; tracking model attention to image, text, instruction, and previously generated tokens, One Token …
The growing demand for image-to-video creation on mobile devices has increasingly focused on cinematic motion effects like bullet time, dolly zoom, slow motion, etc. While Diffusion Transformers (DiTs) exhibit strong performance in video generation, their large parameter sizes and multi-step iterative denoising processes lead to substantial computational overhead, making efficient generation on mobile devices challenging. We propose CineMobile to bridge the gap. In particular, CineMobile adopts…
A key challenge in Arabic NLP is the scarcity of dialectal data relative to Modern Standard Arabic (MSA), causing LLMs to overproduce MSA and struggle with dialectally accurate generation. From an interpretability perspective, this raises a fundamental question: where and how are dialectal features encoded within model internals, and can these representations be leveraged to improve dialect generation without fine-tuning? This study investigates two complementary inference-time approaches that …
We introduce Vidu S1, a real-time interactive video generation model supporting voice control of digital characters. Users can control video generation content at any moment through voice instructions. Vidu S1 supports infinite-length real-time video generation without blurring, drift, or visual distortion. Built with TurboDiffusion and TurboServe, Vidu S1 outputs 540p real-time videos at up to 42 FPS on regular consumer GPUs. Users can upload custom images of real people, anime, and pets, and …