Tagged articles
zero-shot
28 articles · Page 1 of 1
S‑Agent redefines spatial intelligence by replacing single‑shot answers with a multi‑step action chain that combines a vision‑language model for task planning, specialized depth and pose models for 3D alignment, and a spatial expert that converts geometry into usable evidence, achieving state‑of‑the‑art zero‑shot scores on MMSI‑Bench and ViewSpatial‑Bench and further improvements after distilling 29.2 k trajectories into an 8‑billion‑parameter model.
Agent ArchitectureVision-Language Modelmodel distillation
0 likes · 11 min read

The newly open‑sourced Wall‑OSS‑0.5 VLA model demonstrates that a large‑scale pre‑trained embodied robot brain can achieve strong zero‑shot performance on 17 real‑world tasks, exhibit staircase emergence with longer pre‑training, and far surpass the industry baseline after fine‑tuning, while also revealing current precision limits.
Embodied AIPretrainingRobotics
0 likes · 15 min read

FreeOcc introduces a training‑free, open‑vocabulary 3D occupancy prediction framework that combines SLAM‑based pose estimation, 3D Gaussian Splatting, and pretrained vision‑language models to build globally consistent semantic maps, achieving over‑two‑fold IoU improvements on EmbodiedOcc‑ScanNet and strong zero‑shot generalization on the new ReplicaOcc benchmark.
3D GaussianFreeOccSLAM
0 likes · 19 min read

XBridge combines a pre‑trained English‑centric LLM with an external multilingual NMT model via optimal‑transport alignment and a three‑stage training scheme, allowing zero‑training of the LLM while achieving high‑quality question answering and generation for low‑resource and unseen languages, narrowing the performance gap with high‑resource languages.
LLMNMTOptimal Transport
0 likes · 8 min read

The EvoNav framework from Nanjing University of Science and Technology tackles the last‑hundred‑meter challenge of embodied navigation by integrating a Future Chain‑of‑Thought and a Historical Experience chain, achieving significant zero‑shot performance gains on VLN‑CE benchmarks and real‑world robot tests, with code released on GitHub.
Embodied AIEvoNavFuture Chain of Thought
0 likes · 6 min read

The LLM‑in‑Sandbox framework places large language models inside a virtual machine that provides external tool access, persistent storage, and code execution, yielding up to a 24.2% performance boost across six benchmark tasks without additional training, and it scales from zero‑shot to reinforcement‑learning‑enhanced agents while remaining cost‑effective.
Agentic AIEfficiencyLLM
0 likes · 6 min read

This article explains how to converse effectively with large language models by mastering three core prompting techniques—Zero‑shot, Few‑shot, and Chain‑of‑Thought—illustrated with front‑end analogies, code snippets, and a step‑by‑step DeepSeek JSON‑generation exercise that shows common pitfalls and best practices.
DeepSeekJSON generationLLM
0 likes · 12 min read


Amap TechJul 24, 2025 · Artificial Intelligence
The paper introduces FingER, an entity-level evaluation framework and the FingER-Instruct-60k dataset for assessing AI-generated video quality with fine-grained reasoning, and demonstrates state-of-the-art zero-shot performance on multiple benchmarks using novel training strategies.
AI-generated videoDatasetfine-grained evaluation
0 likes · 9 min read

IndexTTS2 introduces a novel auto-regressive zero-shot text-to-speech model that achieves precise duration control and fine-grained emotional expression through a universal time‑encoding mechanism, decoupled voice‑style and emotion modeling, and a GPT‑style latent feature, outperforming state‑of‑the‑art baselines across multiple benchmarks.
Text-to-Speechduration controlemotional synthesis
0 likes · 23 min read


Amap TechJul 9, 2025 · Artificial Intelligence
This article introduces VMBench, the first perception‑aligned video motion generation benchmark that defines a five‑dimensional metric suite and a meta‑guided prompt generation pipeline, and presents LD‑RPS, a zero‑shot unified image restoration framework based on latent diffusion recurrent posterior sampling, together with extensive experiments validating both systems.
benchmarkdiffusion modelsimage restoration
0 likes · 14 min read

This tutorial walks through loading the openai/clip‑vit‑base‑patch32 model with Hugging Face, preprocessing images and text, encoding them into a shared embedding space, computing cosine similarity for zero‑shot image‑text matching, and visualizing the results, all with concrete code examples.
CLIPHugging FacePython
0 likes · 6 min read


AIWalkerMar 13, 2025 · Artificial Intelligence
The paper introduces YOLOE, a new YOLO‑based model that supports text, visual, and no‑prompt open‑world detection and segmentation, detailing its lightweight RepRTA, SAVPE, and LRPC modules and showing benchmark gains in speed and zero‑shot performance on LVIS and COCO.
Computer VisionObject DetectionYOLOE
0 likes · 9 min read


AIWalkerFeb 11, 2025 · Artificial Intelligence
LLMDet introduces a novel training pipeline that leverages large language models to generate detailed image‑level captions and region‑level phrases, fine‑tunes an open‑vocabulary detector with the GroundingCap‑1M dataset, and achieves state‑of‑the‑art zero‑shot performance surpassing Grounding DINO across multiple benchmarks.
GroundingCapLLMDetOpen-Vocabulary Detection
0 likes · 20 min read


AIWalkerFeb 4, 2025 · Artificial Intelligence
The paper introduces MILS, a training‑free multimodal iterative LLM solver that lets large language models perceive and generate across image, video, and audio domains, achieving new state‑of‑the‑art results without any task‑specific data or fine‑tuning.
AI researchLLMMILS
0 likes · 18 min read


AIWalkerJan 15, 2025 · Artificial Intelligence
Magic Mirror introduces a single‑stage, zero‑shot framework that fuses dual facial embeddings with a conditional adaptive normalization module inside a Video Diffusion Transformer, achieving superior identity consistency, natural dynamics, and high visual quality compared with existing video generation methods.
conditional adaptive normalizationdiffusion transformeridentity preservation
0 likes · 16 min read


JD TechNov 12, 2024 · Artificial Intelligence
This article explains what Prompt Engineering is, traces its development from early NLP commands to modern adaptive and multimodal prompting techniques, describes various prompting strategies such as Zero‑shot, Few‑shot, Chain‑of‑Thought, Auto‑CoT, and showcases a JD Logistics case study using these methods to classify product types with code examples.
AI Prompt Designchain-of-thoughtfew-shot
0 likes · 27 min read

InstantID, an open‑source tool released by Xiaohongshu in early 2024, generates multiple stylized portraits that preserve a person’s facial identity from a single reference photo in seconds, eliminating fine‑tuning, large storage needs, and multi‑image requirements while seamlessly working with popular diffusion models like Stable Diffusion 1.5 and SDXL.
AIImage GenerationInstantID
0 likes · 6 min read


AIWalkerJan 22, 2024 · Artificial Intelligence
Depth Anything introduces a highly practical monocular depth estimation model that leverages a 62‑million‑image unlabeled dataset, teacher‑student training, strong data perturbations, and DINOv2‑based semantic supervision to achieve zero‑shot capability and state‑of‑the‑art performance over MiDaS across multiple benchmarks.
Computer VisionDINOv2Monocular Depth
0 likes · 8 min read

This article reviews the evolution of the GPT family from GPT‑1 to GPT‑3, comparing encoder‑decoder architectures, explaining the shift from supervised fine‑tuning to zero‑shot and few‑shot learning, and highlighting the architectural and training innovations that enabled large‑scale language models.
GPTLLMPretraining
0 likes · 13 min read

This comprehensive guide explores prompt engineering for large language models, covering its background, fundamental concepts, prompt formats, construction principles, advanced techniques like few‑shot, zero‑shot, and chain‑of‑thought prompting, as well as practical examples, evaluation metrics, and future directions.
LLMartificial intelligencechain-of-thought
0 likes · 33 min read

This article presents a low‑resource NLP pretraining approach that combines transformer‑based language modeling with contrastive vector learning, details the unsupervised sample‑pair construction, introduces a camel‑shaped masking distribution, and demonstrates through extensive experiments that the resulting model achieves strong zero‑shot NLU, NLG, and retrieval performance while requiring minimal compute and data.
Contrastive LearningPretraininglanguage modeling
0 likes · 10 min read


ByteFEJun 15, 2023 · Artificial Intelligence
This article explains how to craft efficient prompts by combining clear instructions and questions, discusses prompt injection risks and mitigation with delimiters, addresses hallucinations, and introduces zero‑shot, few‑shot, and chain‑of‑thought prompting techniques for large language models.
HallucinationLLMPrompt Injection
0 likes · 16 min read

This article explains how PaddleNLP's Universal Information Extraction (UIE) model can dramatically reduce labeling effort and improve accuracy for logistics parcel data extraction, showcasing a five‑sample experiment that boosts F1 by 18 points to 93% and providing a zero‑shot Python example.
NLPPaddleNLPPython
0 likes · 5 min read

GEN‑VLKT introduces a Guided‑Embedding Network with position‑ and instance‑guided embeddings to remove costly post‑processing and leverages CLIP‑based visual‑linguistic knowledge transfer for interaction understanding, achieving state‑of‑the‑art HOI detection performance and zero‑shot capability, now deployed in Alibaba’s Taobao services.
CLIPHOI detectionTransformer
0 likes · 7 min read

This article explains CLIP’s multimodal contrastive pre‑training, its simple yet effective architecture, code implementation, and how its zero‑shot capability can surpass supervised ImageNet models by leveraging a 400‑million image‑text dataset and shared semantic embeddings.
AICLIPContrastive Learning
0 likes · 7 min read

DeltaLM is a new multilingual pretrained encoder‑decoder model that leverages a pretrained encoder and a novel decoder to improve multilingual neural machine translation, offering efficient training, strong cross‑language transfer, zero‑shot translation, and superior performance on various translation and summarization tasks.
DeltaLMNMTmachine translation
0 likes · 13 min read

DeltaLM is a multilingual pretrained encoder‑decoder model that leverages cross‑lingual transfer from a pretrained encoder and novel decoder architecture, employs span‑corruption and translation‑pair pretraining tasks, and uses a two‑stage fine‑tuning strategy to achieve strong zero‑shot and supervised translation performance across over 100 languages.
Cross-Lingual TransferDeltaLMNeural Machine Translation
0 likes · 12 min read

This article presents Alibaba's comprehensive research on multilingual neural machine translation, covering motivations, model architectures, intermediate language modules, data‑augmentation strategies such as repair translation, integration of pre‑trained models with adapters, and engineering optimizations that enable a production‑ready system supporting over 200 languages.
AdapterAlibabaNeural Machine Translation
0 likes · 21 min read
