ADu2021/skillXiv - Page 29

training-free-group-relative-policy

Improve LLM performance at test time through in-context learning and experience libraries, eliminating the need for parameter updates while maintaining competitive results.

training-free-grpo

Improve LLM outputs without parameter updates using learned token priors that guide inference. Trigger: optimize agent behavior in deployment without model retraining or fine-tuning.

training-x-ray-vision-amodal

Infer complete object structure despite occlusion using multi-camera video. Enables training models to predict hidden object appearance by combining temporal and spatial context from multiple viewpoints.

trajectory-selection-reasoning

Select best reasoning trajectories from multiple samples using step-level scoring from a 0.6B lightweight verifier that exploits hidden states, outperforming external reward models by 4-12% without massive annotations.

trapo-semi-supervised-rl

Bridge labeled and unlabeled data through trajectory similarity in reinforcement learning. Select reliable unlabeled samples by comparing pass-rate evolution trajectories against labeled data. Achieve 42.6% accuracy with 1K labeled + 3K unlabeled samples,…

tree-search-llm-agent-rl

Train LLM agents via tree-search rollout sampling and step-wise advantage estimation. Achieve 1.5x more rollouts within fixed token budgets and implicit step-level preference learning through dual-level advantage computation on tree structures.

treecua-gui-automation

Scale GUI automation by organizing trajectories into tree structures for reuse and branching exploration, reducing data cost while maximizing step-level diversity through adaptive topology.

treegrpo-tree-advantage-rl

Achieve 2.4× faster RL training for diffusion models by restructuring denoising as tree search with shared computation. TreeGRPO computes fine-grained step advantages instead of trajectory-level rewards—crucial for efficient diffusion model optimization.

treevgr-visual-grounded-reasoning-evidence

Train vision-language models to produce visually grounded reasoning by enforcing traceable evidence via bounding box localization, using a novel benchmark (TreeBench) and RL-based training with dual IoU rewards for both recall and precision.

truncated-ppo

T-PPO improves training efficiency via truncated rollouts and extended GAE, enabling batch continuity without waiting for full sequence completion.

truncated-step-level-retrieval-reasoning

Improve credit assignment in retrieval-augmented reasoning by truncating trajectories at single decision points. Generate k samples sharing a common prefix, differing only at the next step to isolate variation and reduce gradient variance by T-fold on T-step…

trust-sql-rl-text-to-sql

Convert natural language to SQL for unknown database schemas by formulating the task as a partially observable MDP. Use dual-track GRPO (token-level masked advantages) to learn schema discovery and query generation jointly.

truth-rl-ternary-reward-hallucination

Train LLMs to reduce hallucinations by 28.9% using a ternary reward scheme that explicitly incentivizes abstention (+0) over false claims (-1) while rewarding correct answers (+1). Apply when improving factual reliability is critical and verification signals…

truthrl-ternary-reward-hallucination-reduction

Reduce LLM hallucinations by training with a ternary reward signal that distinguishes correct answers, hallucinations, and abstentions. This technique incentivizes truthfulness over accuracy-only metrics, enabling safer, more calibrated language models…

ttcs-test-time-curriculum

Improve model performance at test time by dynamically synthesizing curriculum of problem variants. Co-evolving synthesizer and solver agents create reinforcing feedback for continuous improvement without external labels.

ttrv-test-time-rl-vision-language

Adapt vision-language models at test time without labels by extracting implicit reward signals (prediction frequency and entropy) and optimizing via GRPO.

ttt-e2e-long-context

Enable long-context modeling via test-time training with meta-learning. Inner loop continues training on context, compressing information into weights rather than KV cache, outer loop optimizes initialization—maintaining full-attention quality with RNN-like…

tuna-unified-visual-multimodal

Cascaded VAE+SigLIP encoders creating single continuous representation space supporting both vision understanding and generation, trained jointly on both tasks without format mismatches. Deploy for unified multimodal models where understanding and generation…

turbodiffusion-video

Achieve 100-200× video generation speedup via algorithm-system co-optimization. Combines sparse attention acceleration (SageAttention + trainable Sparse-Linear Attention), step distillation, W8A8 quantization, and custom CUDA kernels—maintaining quality…

turn-ppo

Stabilize multi-turn agent RL by shifting from token-level to turn-level MDPs. Reformulates states and actions at conversation-turn granularity, uses learned turn-level critics, and applies Generalized Advantage Estimation—eliminating misalignment that…

tv2tv-interleaved-video-text

Mixture-of-Transformers jointly learning language modeling and video flow matching, enabling interleaved text-video generation where semantic decisions happen in language, pixel generation in video, and users can intervene textually at any step.

twinbrainvla-unleashing-the-potential-of-generalis

Implement techniques from TwinBrainVLA: Unleashing the Potential of Generalist VLMs for Embodied Tasks via Asymmetric Mixture-of-Transformers. The fundamental premise of Vision-Language-Action (VLA) models is to harness the extensive general capabilities of…

twinflow-one-step-generation

Train single-step image generators without teacher models or standard adversarial networks. Achieves 0.83 GenEval score at 1-NFE with 100× computational efficiency gains—when you need real-time image synthesis from pre-trained diffusion models.

two-grpo-contrastive-efficiency

Reduce GRPO training cost by 87.5% using only 2 rollouts instead of 16 while achieving 98.1% of baseline performance. Leverage the insight that GRPO's group mechanism serves contrastive learning rather than advantage estimation.

typhoon-ocr-open-vision-language-model-for-thai

Document extraction is a core component of digital workflows, yet existing vision-language models (VLMs) predominantly favor high-resource languages. Thai presents additional challenges due to script complexity from non-latin letters, the absence of explicit…

u-net-like-spiking-neural-networks-for-single-imag

Research contribution advancing agent and reasoning capabilities through novel approaches to model development, training, and evaluation.

ultramemv2-memory-networks

Scale memory networks to 120B parameters with improved long-context learning through integrated memory layers, simplified value projection, and optimized parameter ratios for superior memory-intensive tasks.

uncertainty-metric-control-signals

Transform uncertainty quantification in LLMs from passive reliability measurement into active control signals for reasoning optimization, autonomous agent decision-making, and reinforcement learning. Use when building systems where uncertainty drives…

unicom-compressed-multimodal-representations

Compress visual embeddings into compact latent space for unified image understanding and generation. Combines attention-based compression with diffusion decoding to bridge comprehension and generation through a shared semantic bottleneck.

unicorn-towards-self-improving-unified-multimodal

Advanced computer vision technique for robust spatial understanding in complex scenes, supporting agent navigation and environmental reasoning capabilities.

unified-thinker-a-general-reasoning-modular-core-f

Advanced reasoning approach for optimizing inference efficiency through meta-cognitive planning, enabling agents to make better decisions with reduced computational overhead.

uniql-quantization-pruning

Single cloud-side weight-sorting and fine-tuning supporting multiple on-device pruning rates via efficient SVD and MLP decomposition, achieving 4-5.7× memory reduction and 2.7-3.4× throughput across Transformers, SSMs, and hybrid architectures.

uniqueness-aware-rl

Improves LLM reasoning by rewarding correct solutions that exhibit rare high-level strategies, preventing exploration collapse and discovering more diverse solution approaches across mathematics, physics, and medical reasoning.

universal-reasoning-model

Enhance Universal Transformers for complex reasoning through ConvSwiGLU modules integrating depthwise convolution into feed-forward blocks and truncated backpropagation through loops (TBPTL) restricting gradient computation to final iterations. Achieve…

uniworld-semantic-vision

Combine semantic encoders from multimodal LLMs with contrastive learning to create unified high-resolution encoders for both visual understanding and generation tasks without relying on VAE compression.

unmasking-diffusion-policies

Learn which tokens to unmask during diffusion sampling via reinforcement learning instead of heuristics. Policies eliminate manual tuning and scale across block sizes—crucial when semi-autoregressive generation needs dynamic, learned unmasking strategies.

unreasonable-scaling-computer-use-agents

Scale computer-use agents from 30% to 72% success rate by generating parallel rollouts and selecting best trajectories through behavior narrative evaluation. Use when deploying desktop agents on complex, high-variance task scenarios.

urbanllava-urban-multimodal-intelligence

Unified MLLM processing four urban data types simultaneously: geospatial structures, trajectory information, satellite imagery, and street-view photos. Outperforms general-purpose models on 12-task urban benchmark with 31-375% improvements. Use for urban…

v-thinker-interactive-visual-reasoning

Enable vision-centric interactive reasoning by synthesizing diverse reasoning datasets through co-evolutionary loops, then training models with progressive curriculum that starts with perception and advances to tool-based problem solving.

va-pi-pixel-ar

Align autoregressive image models with pixel-space quality via variational optimization. Formulates alignment as ELBO combining reconstruction (pixel supervision) and prior regularization (token distribution), treating model as RL policy with tokenizer…