Text understanding and language tasks
We did experiments comparing Optuna & autoresearch. Autoresearch converges faster, is more cost-efficient, and even generalizes bette...
Automate iOS apps with XCUITest and Droidrun using just natural language. You send the command to Droidrun, and the agent starts the task...
I trained a BERT-style transformer on 276K Kubernetes YAML files, replacing standard positional encoding with learned tree coordinates (d...
![[2510.15425] TeamFormer: Shallow Parallel Transformers with Progressive Approximation](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
The paper introduces TeamFormer, a shallow Transformer architecture that enhances parallelism and reduces training time while maintaining...
This article introduces a new paradigm for probabilistic forecasting, proposing 'Probabilistic Scenarios' as an alternative to traditiona...
This paper presents a novel framework for heart rate modeling that addresses data heterogeneity by learning unified representations from ...
The paper introduces One-Step Flow Q-Learning (OFQL), a novel framework that improves offline reinforcement learning by enabling one-step...
The paper presents RefLoRA, a novel method for fine-tuning large models by optimizing low-rank adaptations, leading to faster convergence...
The paper explores the energy-performance tradeoffs in LLM inference across various workloads and GPU scaling, revealing significant insi...
This paper introduces a novel approach to discrete flow matching using minibatch optimal transport, enhancing generative performance whil...
The LUMEN model enhances radiological diagnosis by leveraging longitudinal imaging data and multi-modal training, improving prognostic ca...
The paper introduces SpatiaLQA, a benchmark for evaluating spatial logical reasoning in Vision-Language Models (VLMs), highlighting their...
The paper presents a novel approach to language model distillation by introducing a tail-aware divergence that enhances the influence of ...
This paper investigates the influence of speaker identity on speech spoofing detection systems, proposing a framework that integrates spe...
The paper presents STAR-LDM, a novel language model that integrates latent diffusion planning with autoregressive generation, enhancing n...
The article presents KEMP-PIP, a novel hybrid machine learning framework designed for predicting pro-inflammatory peptides by integrating...
This paper benchmarks distilled language models, demonstrating their superior performance and efficiency in resource-constrained environm...
The paper presents VISION-ICE, an AI framework utilizing intracardiac echocardiography to identify arrhythmia origins, achieving 66.2% ac...
The paper presents UPipe, a novel technique for memory-efficient context parallelism in Transformer models, achieving significant reducti...
The paper presents SOM-VQ, a novel tokenization method that enhances interactive generative models by integrating vector quantization wit...
The article presents ProxyFL, a novel framework for Federated Semi-Supervised Learning (FSSL) that addresses data heterogeneity issues by...
This article evaluates the use of DeepSpeed to enhance the scalability of Vision Transformers (ViTs) for image-centric workloads, focusin...
This article presents a novel approach to EEG-to-text decoding, exploring how hierarchical abstraction levels affect classification perfo...
Get the latest news, tools, and insights delivered to your inbox.
Daily or weekly digest • Unsubscribe anytime