YouZum

Committee

AI, Committee, Noticias, Uncategorized

Energy Landscapes Enable Reliable Abstention in Retrieval-Augmented Large Language Models for Healthcare

arXiv:2509.04482v1 Announce Type: new Abstract: Reliable abstention is critical for retrieval-augmented generation (RAG) systems, particularly in safety-critical domains such as women’s health, where incorrect answers can lead to harm. We present an energy-based model (EBM) that learns a smooth energy landscape over a dense semantic corpus of 2.6M guideline-derived questions, enabling the system to decide when to generate or abstain. We benchmark the EBM against a calibrated softmax baseline and a k-nearest neighbour (kNN) density heuristic across both easy and hard abstention splits, where hard cases are semantically challenging near-distribution queries. The EBM achieves superior abstention performance abstention on semantically hard cases, reaching AUROC 0.961 versus 0.950 for softmax, while also reducing FPR@95 (0.235 vs 0.331). On easy negatives, performance is comparable across methods, but the EBM’s advantage becomes most pronounced in safety-critical hard distributions. A comprehensive ablation with controlled negative sampling and fair data exposure shows that robustness stems primarily from the energy scoring head, while the inclusion or exclusion of specific negative types (hard, easy, mixed) sharpens decision boundaries but is not essential for generalisation to hard cases. These results demonstrate that energy-based abstention scoring offers a more reliable confidence signal than probability-based softmax confidence, providing a scalable and interpretable foundation for safe RAG systems.

Energy Landscapes Enable Reliable Abstention in Retrieval-Augmented Large Language Models for Healthcare Leer entrada »

AI, Committee, Noticias, Uncategorized

Research on Multi-hop Inference Optimization of LLM Based on MQUAKE Framework

arXiv:2509.04770v1 Announce Type: new Abstract: Accurately answering complex questions has consistently been a significant challenge for Large Language Models (LLMs). To address this, this paper proposes a multi-hop question decomposition method for complex questions, building upon research within the MQUAKE framework. Utilizing the LLAMA3 model, we systematically investigate the impact of multi-hop question decomposition within knowledge graphs on model comprehension and reasoning accuracy, both before and after model training. In our experiments, we systematically partitioned and converted the MQUAKE-T dataset into two distinct formats: a single-hop dataset designed for directly answering complex questions, and a multi-hop dataset constructed using the multi-hop question decomposition method. We then fine-tuned the LLAMA3 model on these datasets and conducted inference tests. Our results demonstrate that, without fine-tuning the LLM, the prediction performance based on the multi-hop question decomposition method significantly outperforms the method of directly answering complex questions. After fine-tuning using the LoRA (Low-Rank Adaptation) method, the performance of both approaches improved compared to the untrained baseline. Crucially, the method utilizing multi-hop decomposition consistently maintained its superiority. These findings validate the effectiveness of the multi-hop decomposition method both before and after training, demonstrating its capability to effectively enhance the LLM’s ability to answer complex questions.

Research on Multi-hop Inference Optimization of LLM Based on MQUAKE Framework Leer entrada »

AI, Committee, Noticias, Uncategorized

DecMetrics: Structured Claim Decomposition Scoring for Factually Consistent LLM Outputs

arXiv:2509.04483v1 Announce Type: new Abstract: Claim decomposition plays a crucial role in the fact-checking process by breaking down complex claims into simpler atomic components and identifying their unfactual elements. Despite its importance, current research primarily focuses on generative methods for decomposition, with insufficient emphasis on evaluating the quality of these decomposed atomic claims. To bridge this gap, we introduce textbf{DecMetrics}, which comprises three new metrics: texttt{COMPLETENESS}, texttt{CORRECTNESS}, and texttt{SEMANTIC ENTROPY}, designed to automatically assess the quality of claims produced by decomposition models. Utilizing these metrics, we develop a lightweight claim decomposition model, optimizing its performance through the integration of these metrics as a reward function. Through automatic evaluation, our approach aims to set a benchmark for claim decomposition, enhancing both the reliability and effectiveness of fact-checking systems.

DecMetrics: Structured Claim Decomposition Scoring for Factually Consistent LLM Outputs Leer entrada »

AI, Committee, Noticias, Uncategorized

Entropy2Vec: Crosslingual Language Modeling Entropy as End-to-End Learnable Language Representations

arXiv:2509.05060v1 Announce Type: new Abstract: We introduce Entropy2Vec, a novel framework for deriving cross-lingual language representations by leveraging the entropy of monolingual language models. Unlike traditional typological inventories that suffer from feature sparsity and static snapshots, Entropy2Vec uses the inherent uncertainty in language models to capture typological relationships between languages. By training a language model on a single language, we hypothesize that the entropy of its predictions reflects its structural similarity to other languages: Low entropy indicates high similarity, while high entropy suggests greater divergence. This approach yields dense, non-sparse language embeddings that are adaptable to different timeframes and free from missing values. Empirical evaluations demonstrate that Entropy2Vec embeddings align with established typological categories and achieved competitive performance in downstream multilingual NLP tasks, such as those addressed by the LinguAlchemy framework.

Entropy2Vec: Crosslingual Language Modeling Entropy as End-to-End Learnable Language Representations Leer entrada »

AI, Committee, Noticias, Uncategorized

Language-Driven Hierarchical Task Structures as Explicit World Models for Multi-Agent Learning

arXiv:2509.04731v1 Announce Type: cross Abstract: The convergence of Language models, Agent models, and World models represents a critical frontier for artificial intelligence. While recent progress has focused on scaling Language and Agent models, the development of sophisticated, explicit World Models remains a key bottleneck, particularly for complex, long-horizon multi-agent tasks. In domains such as robotic soccer, agents trained via standard reinforcement learning in high-fidelity but structurally-flat simulators often fail due to intractable exploration spaces and sparse rewards. This position paper argues that the next frontier in developing capable agents lies in creating environments that possess an explicit, hierarchical World Model. We contend that this is best achieved through hierarchical scaffolding, where complex goals are decomposed into structured, manageable subgoals. Drawing evidence from a systematic review of 2024 research in multi-agent soccer, we identify a clear and decisive trend towards integrating symbolic and hierarchical methods with multi-agent reinforcement learning (MARL). These approaches implicitly or explicitly construct a task-based world model to guide agent learning. We then propose a paradigm shift: leveraging Large Language Models to dynamically generate this hierarchical scaffold, effectively using language to structure the World Model on the fly. This language-driven world model provides an intrinsic curriculum, dense and meaningful learning signals, and a framework for compositional learning, enabling Agent Models to acquire sophisticated, strategic behaviors with far greater sample efficiency. By building environments with explicit, language-configurable task layers, we can bridge the gap between low-level reactive behaviors and high-level strategic team play, creating a powerful and generalizable framework for training the next generation of intelligent agents.

Language-Driven Hierarchical Task Structures as Explicit World Models for Multi-Agent Learning Leer entrada »

AI, Committee, Noticias, Uncategorized

Alibaba AI Unveils Qwen3-Max Preview: A Trillion-Parameter Qwen Model with Super Fast Speed and Quality

Alibaba’s Qwen Team unveiled Qwen3-Max-Preview (Instruct), a new flagship large language model with over one trillion parameters—their largest to date. It is accessible through Qwen Chat, Alibaba Cloud API, OpenRouter, and as default in Hugging Face’s AnyCoder tool. How does it fit in today’s LLM landscape? This milestone comes at a time when the industry is trending toward smaller, more efficient models. Alibaba’s decision to move upward in scale marks a deliberate strategic choice, highlighting both its technical capabilities and commitment to trillion-parameter research. How large is Qwen3-Max and what are its context limits? Parameters: >1 trillion. Context window: Up to 262,144 tokens (258,048 input, 32,768 output). Efficiency feature: Includes context caching to speed up multi-turn sessions. How does Qwen3-Max perform against other models? Benchmarks show it outperforms Qwen3-235B-A22B-2507 and competes strongly with Claude Opus 4, Kimi K2, and Deepseek-V3.1 across SuperGPQA, AIME25, LiveCodeBench v6, Arena-Hard v2, and LiveBench. What is the pricing structure for usage? Alibaba Cloud applies tiered token-based pricing: 0–32K tokens: $0.861/million input, $3.441/million output 32K–128K: $1.434/million input, $5.735/million output 128K–252K: $2.151/million input, $8.602/million output This model is cost-efficient for smaller tasks but scales up significantly in price for long-context workloads. How does the closed-source approach impact adoption? Unlike earlier Qwen releases, this model is not open-weight. Access is restricted to APIs and partner platforms. This choice highlights Alibaba’s commercialization focus but may slow broader adoption in research and open-source communities Key Takeaways First trillion-parameter Qwen model – Qwen3-Max surpasses 1T parameters, making it Alibaba’s largest and most advanced LLM to date. Ultra-long context handling – Supports 262K tokens with caching, enabling extended document and session processing beyond most commercial models. Competitive benchmark performance – Outperforms Qwen3-235B and competes with Claude Opus 4, Kimi K2, and Deepseek-V3.1 on reasoning, coding, and general tasks. Emergent reasoning despite design – Though not marketed as a reasoning model, early results show structured reasoning capabilities on complex tasks. Closed-source, tiered pricing model – Available via APIs with token-based pricing; economical for small tasks but costly at higher context usage, limiting accessibility. Summary Qwen3-Max-Preview sets a new scale benchmark in commercial LLMs. Its trillion-parameter design, 262K context length, and strong benchmark results highlight Alibaba’s technical depth. Yet the model’s closed-source release and steep tiered pricing create a question for broader accessibility. Check out the Qwen Chat and Alibaba Cloud API. Feel free to check out our GitHub Page for Tutorials, Codes and Notebooks. Also, feel free to follow us on Twitter and don’t forget to join our 100k+ ML SubReddit and Subscribe to our Newsletter. The post Alibaba AI Unveils Qwen3-Max Preview: A Trillion-Parameter Qwen Model with Super Fast Speed and Quality appeared first on MarkTechPost.

Alibaba AI Unveils Qwen3-Max Preview: A Trillion-Parameter Qwen Model with Super Fast Speed and Quality Leer entrada »

AI, Committee, Noticias, Uncategorized

Implementing DeepSpeed for Scalable Transformers: Advanced Training with Gradient Checkpointing and Parallelism

In this advanced DeepSpeed tutorial, we provide a hands-on walkthrough of cutting-edge optimization techniques for training large language models efficiently. By combining ZeRO optimization, mixed-precision training, gradient accumulation, and advanced DeepSpeed configurations, the tutorial demonstrates how to maximize GPU memory utilization, reduce training overhead, and enable scaling of transformer models in resource-constrained environments, such as Colab. Alongside model creation and training, it also covers performance monitoring, inference optimization, checkpointing, and benchmarking different ZeRO stages, providing practitioners with both theoretical insights and practical code to accelerate model development. Check out the FULL CODES here. Copy CodeCopiedUse a different Browser import subprocess import sys import os import json import time from pathlib import Path def install_dependencies(): “””Install required packages for DeepSpeed in Colab””” print(” Installing DeepSpeed and dependencies…”) subprocess.check_call([ sys.executable, “-m”, “pip”, “install”, “torch”, “torchvision”, “torchaudio”, “–index-url”, “https://download.pytorch.org/whl/cu118” ]) subprocess.check_call([sys.executable, “-m”, “pip”, “install”, “deepspeed”]) subprocess.check_call([ sys.executable, “-m”, “pip”, “install”, “transformers”, “datasets”, “accelerate”, “wandb” ]) print(” Installation complete!”) install_dependencies() import torch import torch.nn as nn import torch.optim as optim from torch.utils.data import DataLoader, Dataset import deepspeed from transformers import GPT2Config, GPT2LMHeadModel, GPT2Tokenizer import numpy as np from typing import Dict, Any import argparse We set up our Colab environment by installing PyTorch with CUDA support, DeepSpeed, and essential libraries like Transformers, Datasets, Accelerate, and Weights & Biases. We ensure everything is ready so we can smoothly build and train models with DeepSpeed. Check out the FULL CODES here. Copy CodeCopiedUse a different Browser class SyntheticTextDataset(Dataset): “””Synthetic dataset for demonstration purposes””” def __init__(self, size: int = 1000, seq_length: int = 512, vocab_size: int = 50257): self.size = size self.seq_length = seq_length self.vocab_size = vocab_size self.data = torch.randint(0, vocab_size, (size, seq_length)) def __len__(self): return self.size def __getitem__(self, idx): return { ‘input_ids’: self.data[idx], ‘labels’: self.data[idx].clone() } We create a SyntheticTextDataset where we generate random token sequences to mimic real text data. We use these sequences as both inputs and labels, allowing us to quickly test DeepSpeed training without relying on a large external dataset. Check out the FULL CODES here. Copy CodeCopiedUse a different Browser class AdvancedDeepSpeedTrainer: “””Advanced DeepSpeed trainer with multiple optimization techniques””” def __init__(self, model_config: Dict[str, Any], ds_config: Dict[str, Any]): self.model_config = model_config self.ds_config = ds_config self.model = None self.engine = None self.tokenizer = None def create_model(self): “””Create a GPT-2 style model for demonstration””” print(” Creating model…”) config = GPT2Config( vocab_size=self.model_config[‘vocab_size’], n_positions=self.model_config[‘seq_length’], n_embd=self.model_config[‘hidden_size’], n_layer=self.model_config[‘num_layers’], n_head=self.model_config[‘num_heads’], resid_pdrop=0.1, embd_pdrop=0.1, attn_pdrop=0.1, ) self.model = GPT2LMHeadModel(config) self.tokenizer = GPT2Tokenizer.from_pretrained(‘gpt2’) self.tokenizer.pad_token = self.tokenizer.eos_token print(f” Model parameters: {sum(p.numel() for p in self.model.parameters()):,}”) return self.model def create_deepspeed_config(self): “””Create comprehensive DeepSpeed configuration””” return { “train_batch_size”: self.ds_config[‘train_batch_size’], “train_micro_batch_size_per_gpu”: self.ds_config[‘micro_batch_size’], “gradient_accumulation_steps”: self.ds_config[‘gradient_accumulation_steps’], “zero_optimization”: { “stage”: self.ds_config[‘zero_stage’], “allgather_partitions”: True, “allgather_bucket_size”: 5e8, “overlap_comm”: True, “reduce_scatter”: True, “reduce_bucket_size”: 5e8, “contiguous_gradients”: True, “cpu_offload”: self.ds_config.get(‘cpu_offload’, False) }, “fp16”: { “enabled”: True, “loss_scale”: 0, “loss_scale_window”: 1000, “initial_scale_power”: 16, “hysteresis”: 2, “min_loss_scale”: 1 }, “optimizer”: { “type”: “AdamW”, “params”: { “lr”: self.ds_config[‘learning_rate’], “betas”: [0.9, 0.999], “eps”: 1e-8, “weight_decay”: 0.01 } }, “scheduler”: { “type”: “WarmupLR”, “params”: { “warmup_min_lr”: 0, “warmup_max_lr”: self.ds_config[‘learning_rate’], “warmup_num_steps”: 100 } }, “gradient_clipping”: 1.0, “wall_clock_breakdown”: True, “memory_breakdown”: True, “tensorboard”: { “enabled”: True, “output_path”: “./logs/”, “job_name”: “deepspeed_advanced_tutorial” } } def initialize_deepspeed(self): “””Initialize DeepSpeed engine””” print(” Initializing DeepSpeed…”) parser = argparse.ArgumentParser() parser.add_argument(‘–local_rank’, type=int, default=0) args = parser.parse_args([]) self.engine, optimizer, _, lr_scheduler = deepspeed.initialize( args=args, model=self.model, config=self.create_deepspeed_config() ) print(f” DeepSpeed engine initialized with ZeRO stage {self.ds_config[‘zero_stage’]}”) return self.engine def train_step(self, batch: Dict[str, torch.Tensor]) -> Dict[str, float]: “””Perform a single training step with DeepSpeed optimizations””” input_ids = batch[‘input_ids’].to(self.engine.device) labels = batch[‘labels’].to(self.engine.device) outputs = self.engine(input_ids=input_ids, labels=labels) loss = outputs.loss self.engine.backward(loss) self.engine.step() return { ‘loss’: loss.item(), ‘lr’: self.engine.lr_scheduler.get_last_lr()[0] if self.engine.lr_scheduler else 0 } def train(self, dataloader: DataLoader, num_epochs: int = 2): “””Complete training loop with monitoring””” print(f” Starting training for {num_epochs} epochs…”) self.engine.train() total_steps = 0 for epoch in range(num_epochs): epoch_loss = 0.0 epoch_steps = 0 print(f”n Epoch {epoch + 1}/{num_epochs}”) for step, batch in enumerate(dataloader): start_time = time.time() metrics = self.train_step(batch) epoch_loss += metrics[‘loss’] epoch_steps += 1 total_steps += 1 if step % 10 == 0: step_time = time.time() – start_time print(f” Step {step:4d} | Loss: {metrics[‘loss’]:.4f} | ” f”LR: {metrics[‘lr’]:.2e} | Time: {step_time:.3f}s”) if step % 20 == 0 and hasattr(self.engine, ‘monitor’): self.log_memory_stats() if step >= 50: break avg_loss = epoch_loss / epoch_steps print(f” Epoch {epoch + 1} completed | Average Loss: {avg_loss:.4f}”) print(” Training completed!”) def log_memory_stats(self): “””Log GPU memory statistics””” if torch.cuda.is_available(): allocated = torch.cuda.memory_allocated() / 1024**3 reserved = torch.cuda.memory_reserved() / 1024**3 print(f” GPU Memory – Allocated: {allocated:.2f}GB | Reserved: {reserved:.2f}GB”) def save_checkpoint(self, path: str): “””Save model checkpoint using DeepSpeed””” print(f” Saving checkpoint to {path}”) self.engine.save_checkpoint(path) def demonstrate_inference(self, text: str = “The future of AI is”): “””Demonstrate optimized inference with DeepSpeed””” print(f”n Running inference with prompt: ‘{text}'”) inputs = self.tokenizer.encode(text, return_tensors=’pt’).to(self.engine.device) self.engine.eval() with torch.no_grad(): outputs = self.engine.module.generate( inputs, max_length=inputs.shape[1] + 50, num_return_sequences=1, temperature=0.8, do_sample=True, pad_token_id=self.tokenizer.eos_token_id ) generated_text = self.tokenizer.decode(outputs[0], skip_special_tokens=True) print(f” Generated text: {generated_text}”) self.engine.train() We build an end-to-end trainer that creates a GPT-2 model, sets a DeepSpeed config (ZeRO, FP16, AdamW, warmup scheduler, tensorboard), and initializes the engine. We then run efficient training steps with logging and memory statistics, save checkpoints, and demonstrate inference to verify optimization and generation in one place. Check out the FULL CODES here. Copy CodeCopiedUse a different Browser def run_advanced_tutorial(): “””Main function to run the advanced DeepSpeed tutorial””” print(” Advanced DeepSpeed Tutorial Starting…”) print(“=” * 60) model_config = { ‘vocab_size’: 50257, ‘seq_length’: 512, ‘hidden_size’: 768, ‘num_layers’: 6, ‘num_heads’: 12 } ds_config = { ‘train_batch_size’: 16, ‘micro_batch_size’: 4, ‘gradient_accumulation_steps’: 4, ‘zero_stage’: 2, ‘learning_rate’: 1e-4, ‘cpu_offload’: False } print(” Configuration:”) print(f” Model size: ~{sum(np.prod(shape) for shape in [[model_config[‘vocab_size’], model_config[‘hidden_size’]], [model_config[‘hidden_size’], model_config[‘hidden_size’]] * model_config[‘num_layers’]]) / 1e6:.1f}M parameters”) print(f” ZeRO Stage: {ds_config[‘zero_stage’]}”) print(f” Batch size: {ds_config[‘train_batch_size’]}”) trainer = AdvancedDeepSpeedTrainer(model_config, ds_config) model = trainer.create_model() engine = trainer.initialize_deepspeed() print(“n Creating synthetic dataset…”) dataset = SyntheticTextDataset( size=200, seq_length=model_config[‘seq_length’], vocab_size=model_config[‘vocab_size’] ) dataloader = DataLoader( dataset, batch_size=ds_config[‘micro_batch_size’], shuffle=True ) print(“n Pre-training memory stats:”) trainer.log_memory_stats() trainer.train(dataloader, num_epochs=2) print(“n Post-training memory stats:”) trainer.log_memory_stats() trainer.demonstrate_inference(“DeepSpeed enables efficient training of”) checkpoint_path = “./deepspeed_checkpoint” trainer.save_checkpoint(checkpoint_path) demonstrate_zero_stages() demonstrate_memory_optimization() print(“n Tutorial completed successfully!”) print(“Key DeepSpeed features demonstrated:”) print(” ZeRO optimization for memory efficiency”) print(” Mixed precision training (FP16)”) print(” Gradient accumulation”) print(” Learning

Implementing DeepSpeed for Scalable Transformers: Advanced Training with Gradient Checkpointing and Parallelism Leer entrada »

AI, Committee, Noticias, Uncategorized

Hugging Face Open-Sourced FineVision: A New Multimodal Dataset with 24 Million Samples for Training Vision-Language Models (VLMs)

Hugging Face has just released FineVision, an open multimodal dataset designed to set a new standard for Vision-Language Models (VLMs). With 17.3 million images, 24.3 million samples, 88.9 million question-answer turns, and nearly 10 billion answer tokens, FineVision position itself as one of the largest and structured publicly available VLM training datasets. FineVision aggregates 200+ sources into a unified format, rigorously filtered for duplicates and benchmark contamination. Rated systematically across multiple quality dimensions, the dataset enables researchers and devs to construct robust training mixtures while minimizing data leakage. Why is FineVision Important for VLM Training? Most state-of-the-art VLMs rely on proprietary datasets, limiting reproducibility and accessibility for the broader research community. FineVision addresses this gap by: Scale and Coverage: 5 TB of curated data across 9 categories, including General VQA, OCR QA, Chart & Table reasoning, Science, Captioning, Grounding & Counting, and GUI navigation. Benchmark Gains: Across 11 widely used benchmarks (e.g., AI2D, ChartQA, DocVQA, ScienceQA, OCRBench), models trained on FineVision outperform alternatives by significant margins—up to 46.3% over LLaVA, 40.7% over Cauldron, and 12.1% over Cambrian. New Skill Domains: FineVision introduces data for emerging tasks like GUI navigation, pointing, and counting, expanding the capabilities of VLMs beyond conventional captioning and VQA. How Was FineVision Built? The curation pipeline followed a three-step process: Collection and AugmentationOver 200 publicly available image-text datasets were gathered. Missing modalities (e.g., text-only data) were reformatted into QA pairs. Underrepresented domains, such as GUI data, were supplemented through targeted collection. Cleaning Removed oversized QA pairs (>8192 tokens). Resized large images to a maximum of 2048 px while preserving aspect ratio. Discarded corrupted samples. Quality RatingUsing Qwen3-32B and Qwen2.5-VL-32B-Instruct as judges, every QA pair was rated on four axes: Text Formatting Quality Question-Answer Relevance Visual Dependency Image-Question Correspondence These ratings enable selective training mixtures, though ablations show that retaining all samples yields the best performance, even when lower-rated samples are included. Comparative Analysis: FineVision vs. Existing Open Datasets Dataset Images Samples Turns Tokens Leakage Perf. Drop After Deduplication Cauldron 2.0M 1.8M 27.8M 0.3B 3.05% -2.39% LLaVA-Vision 2.5M 3.9M 9.1M 1.0B 2.15% -2.72% Cambrian-7M 5.4M 7.0M 12.2M 0.8B 2.29% -2.78% FineVision 17.3M 24.3M 88.9M 9.5B 1.02% -1.45% FineVision is not only one of the largest but also the least hallucinated dataset, with just 1% overlap with benchmark test sets. This ensures minimal data leakage and reliable evaluation performance. Performance Insights Model Setup: Ablations were conducted using nanoVLM (460M parameters), combining SmolLM2-360M-Instruct as the language backbone and SigLIP2-Base-512 as the vision encoder. Training Efficiency: On 32 NVIDIA H100 GPUs, one full epoch (12k steps) takes ~20 hours. Performance Trends: FineVision models improve steadily with exposure to diverse data, overtaking baselines after ~12k steps. Deduplication experiments confirm FineVision’s low leakage compared to Cauldron, LLaVA, and Cambrian. Multilingual subsets, even when the backbone is monolingual, show slight performance gains, suggesting diversity outweighs strict alignment. Attempts at multi-stage training (two or 2.5 stages) did not yield consistent benefits, reinforcing that scale + diversity is more critical than training heuristics. Why FineVision Brings the New Standard? +20% Average Performance Boost: Outperforms all existing open datasets across 10+ benchmarks. Unprecedented Scale: 17M+ images, 24M+ samples, 10B tokens. Skill Expansion: GUI navigation, counting, pointing, and document reasoning included. Lowest Data Leakage: 1% contamination, compared to 2–3% in other datasets. Fully Open Source: Available on Hugging Face Hub for immediate use via the datasets library. Conclusion FineVision marks a significant advancement in open multimodal datasets. Its large scale, systematic curation, and transparent quality assessments create a reproducible and extensible foundation for training state-of-the-art Vision-Language Models. By reducing dependence on proprietary resources, it enables researchers and devs to build competitive systems and accelerate progress in areas such as document analysis, visual reasoning, and agentic multimodal tasks. Check out the Dataset and Technical details. Feel free to check out our GitHub Page for Tutorials, Codes and Notebooks. Also, feel free to follow us on Twitter and don’t forget to join our 100k+ ML SubReddit and Subscribe to our Newsletter. The post Hugging Face Open-Sourced FineVision: A New Multimodal Dataset with 24 Million Samples for Training Vision-Language Models (VLMs) appeared first on MarkTechPost.

Hugging Face Open-Sourced FineVision: A New Multimodal Dataset with 24 Million Samples for Training Vision-Language Models (VLMs) Leer entrada »

AI, Committee, Noticias, Uncategorized

Tilde AI Releases TildeOpen LLM: An Open-Source Large Language Model with Over 30 Billion Parameters and Support Most European Languages

Latvian language-tech firm Tilde has released TildeOpen LLM, an open-source foundational large language model (LLM) purpose-built for European languages, with a sharp focus on under-represented and smaller national and regional languages. It’s a strategic leap toward linguistic equity and digital sovereignty within the EU. Under the Hood: Architecture, Training and Governance The public release occurred on September 3, 2025, when Tilde deployed the model free to users via Hugging Face. Built as a 30-billion-parameter dense decoder-only transformer, the model is available under a permissive license (CC-BY-4.0) and includes broad language support—from Latvian and Lithuanian to Ukrainian, Turkish, and beyond. Training occurred on the EU’s supercomputers: LUMI (Finland) and JUPITER, tapping into 2 million GPU hours awarded via the European Commission’s Large AI Grand Challenge. Fine technical detail: trained via EleutherAI–inspired GPT-NeoX scripts across 450K updates, consuming ~2 trillion tokens. Training included three-stage sampling: uniform across languages, natural distribution to boost high-data-volume languages, and a final uniform sweep for balance. Hyperparameters: 60 layers, embedding size 6144, 48 attention heads, 8192-token context window, SwiGLU activations, RoPE positional encoding, RMSNorm layer norms. Language Equity and Data Sovereignty Mainstream models lean heavily on English and other major languages, causing skewed performance when dealing with Baltic, Slavic, or other smaller European languages. This under-representation leads to poor grammar, awkward phrasing, and hallucinations. TildeOpen resolves this by embedding an “equitable tokenizer”, engineered to represent text similarly regardless of language—reducing token count and increasing inference efficiency for lesser-represented languages. Crucially, organizations can self-host—in local data centers or secure EU-compliant clouds—ensuring adherence to GDPR and other data-protection mandates. This addresses sovereignty concerns tied to US- or Asia-hosted models. Strategic Horizon: From Prototype to European AI Infrastructure TildeOpen is a foundational “base” model. It is expected for it’s upcoming versions more specialized (e.g., instruction-tuned translation models) built atop this core. It’s also a geo-flag planting moment: Latvia, via Tilde, positions itself as a tech exporter, with aspirations to scale European AI infrastructure while preserving linguistic diversity. For Research, the move mirrors broader research on multilingual model behavior—gaps still exist. Evaluations show even strong open LLMs can hallucinate or lag in lexical accuracy for Baltic languages, reinforcing the need for localized development. Summary TildeOpen LLM reframes EU AI—not just as regulatory compliance, but as technical stewardship. It’s a grounded, high-capacity model with transparent architecture, scalable deployment, and a fierce commitment to linguistic equity. It doesn’t indulge hype; it delivers substance. FAQs Q1: What is TildeOpen LLM?TildeOpen is a 30B-parameter multilingual large language model trained on EU supercomputers, optimized for European languages, especially under-represented ones. Q2: How is it different from mainstream LLMs?Unlike global models that prioritize English, TildeOpen uses an equitable tokenizer and balanced training to ensure fair representation and accuracy across smaller European languages. Q3: Can organizations self-host the model?Yes. TildeOpen is open-source under CC-BY-4.0 and can be deployed in local data centers or EU-compliant clouds to meet GDPR and data sovereignty requirements. Q4: What are the main use cases?Government services, translation, education, AI assistants, speech technologies, and multilingual customer support—any domain requiring accurate European language processing. Check out the Model on Hugging Face and Technical details here. Feel free to check out our GitHub Page for Tutorials, Codes and Notebooks. Also, feel free to follow us on Twitter and don’t forget to join our 100k+ ML SubReddit and Subscribe to our Newsletter. The post Tilde AI Releases TildeOpen LLM: An Open-Source Large Language Model with Over 30 Billion Parameters and Support Most European Languages appeared first on MarkTechPost.

Tilde AI Releases TildeOpen LLM: An Open-Source Large Language Model with Over 30 Billion Parameters and Support Most European Languages Leer entrada »

AI, Committee, Noticias, Uncategorized

From Pretraining to Post-Training: Why Language Models Hallucinate and How Evaluation Methods Reinforce the Problem

Large language models (LLMs) very often generate “hallucinations”—confident yet incorrect outputs that appear plausible. Despite improvements in training methods and architectures, hallucinations persist. A new research from OpenAI provides a rigorous explanation: hallucinations stem from statistical properties of supervised versus self-supervised learning, and their persistence is reinforced by misaligned evaluation benchmarks. What Makes Hallucinations Statistically Inevitable? The research team explains hallucinations as errors inherent to generative modeling. Even with perfectly clean training data, the cross-entropy objective used in pretraining introduces statistical pressures that produce errors. The research team reduce the problem to a supervised binary classification task called Is-It-Valid (IIV): determining whether a model’s output is valid or erroneous. They prove that the generative error rate of an LLM is at least twice its IIV misclassification rate. In other words, hallucinations occur for the same reasons misclassifications appear in supervised learning: epistemic uncertainty, poor models, distribution shift, or noisy data. Why Do Rare Facts Trigger More Hallucinations? One major driver is the singleton rate—the fraction of facts that appear only once in training data. By analogy to Good–Turing missing-mass estimation, if 20% of facts are singletons, at least 20% of them will be hallucinated. This explains why LLMs answer reliably about widely repeated facts (e.g., Einstein’s birthday) but fail on obscure or rarely mentioned ones. Can Poor Model Families Lead to Hallucinations? Yes. Hallucinations also emerge when the model class cannot adequately represent a pattern. Classic examples include n-gram models generating ungrammatical sentences, or modern tokenized models miscounting letters because characters are hidden inside subword tokens. These representational limits cause systematic errors even when the data itself is sufficient. Why Doesn’t Post-Training Eliminate Hallucinations? Post-training methods such as RLHF (reinforcement learning from human feedback), DPO, and RLAIF reduce some errors, especially harmful or conspiratorial outputs. But overconfident hallucinations remain because evaluation incentives are misaligned. Like students guessing on multiple-choice exams, LLMs are rewarded for bluffing when unsure. Most benchmarks—such as MMLU, GPQA, and SWE-bench—apply binary scoring: correct answers get credit, abstentions (“I don’t know”) get none, and incorrect answers are penalized no more harshly than abstentions. Under this scheme, guessing maximizes benchmark scores, even if it fosters hallucinations. How Do Leaderboards Reinforce Hallucinations? A review of popular benchmarks shows that nearly all use binary grading with no partial credit for uncertainty. As a result, models that truthfully express uncertainty perform worse than those that always guess. This creates systemic pressure for developers to optimize models for confident answers rather than calibrated ones. What Changes Could Reduce Hallucinations? The research team argue that fixing hallucinations requires socio-technical change, not just new evaluation suites. They propose explicit confidence targets: benchmarks should clearly specify penalties for wrong answers and partial credit for abstentions. For example: “Answer only if you are >75% confident. Mistakes lose 2 points; correct answers earn 1; ‘I don’t know’ earns 0.” This design mirrors real-world exams like earlier SAT and GRE formats, where guessing carried penalties. It encourages behavioral calibration—models abstain when their confidence is below the threshold, producing fewer overconfident hallucinations while still optimizing for benchmark performance. What Are the Broader Implications? This work reframes hallucinations as predictable outcomes of training objectives and evaluation misalignment rather than inexplicable quirks. The findings highlight: Pretraining inevitability: Hallucinations parallel misclassification errors in supervised learning. Post-training reinforcement: Binary grading schemes incentivize guessing. Evaluation reform: Adjusting mainstream benchmarks to reward uncertainty can realign incentives and improve trustworthiness. By connecting hallucinations to established learning theory, the research demystifies their origin and suggests practical mitigation strategies that shift responsibility from model architectures to evaluation design. Check out the PAPER and Technical details here. Feel free to check out our GitHub Page for Tutorials, Codes and Notebooks. Also, feel free to follow us on Twitter and don’t forget to join our 100k+ ML SubReddit and Subscribe to our Newsletter. The post From Pretraining to Post-Training: Why Language Models Hallucinate and How Evaluation Methods Reinforce the Problem appeared first on MarkTechPost.

From Pretraining to Post-Training: Why Language Models Hallucinate and How Evaluation Methods Reinforce the Problem Leer entrada »

We use cookies to improve your experience and performance on our website. You can learn more at Política de privacidad and manage your privacy settings by clicking Settings.

Privacy Preferences

You can choose your cookie settings by turning on/off each type of cookie as you wish, except for essential cookies.

Allow All
Manage Consent Preferences
  • Always Active

Save
es_ES