arXiv subjects

Francisco Guzmán

Publications and source records attributed to Francisco Guzmán.

At least 19 recordsLinked to original sources

arXiv2026

Divergent Paths: Learning an Outcome-Predictive Representation of LLM Agent Trajectories

In this work, we reframe agentic trajectories as paths through a shared embedding space of actions and environmental observations. We show that successful and failed rollouts of the same task separate in this space, with simple per-task discriminators predicting held-out rollout outcomes at up to 0.95 AUROC across three agent benchmarks and three policy models. To exploit this signal across tasks, we introduce Precog-RM, a transformer-based model that meta-learns a representation in which outcome prediction can be performed in a few-shot setting. We show that Precog-RM generalizes to unseen tasks and models, and gives confident predictions of trajectory success long before termination. Deployed as an early stopping controller for Group Relative Policy Optimization (GRPO) on Tau2-Bench, Precog-RM cuts training FLOPs by up to 47.3\% while maintaining performance comparable to baseline GRPO.

cs.LG↗

arXiv2026

An Imperfect Verifier is Good Enough: Learning with Noisy Rewards

Reinforcement Learning with Verifiable Rewards (RLVR) is widely used for post-training Large Language Models, but practical verifiers can make errors. We study how the rate and structure of reward noise affect RLVR in code generation, with a preliminary scientific-reasoning check. In multi-seed Qwen3 8B experiments on MBPP, mean validation reward over two fixed late evaluations is within 1 percentage point of the clean baseline at the tested resampled group-rollout noise rates through 20%, and within about 2 points at 30%. Confidence intervals allow larger losses; these point estimates do not establish a general tolerance threshold. We also examine full-program pass@k, four controlled noise structures, two model-based verifiers, and policy models from three families spanning 4B-9B parameters. We derive conditional advantage distributions for symmetric and asymmetric group noise, including retained format penalties, and show why clipping limits simple gradient-scaling arguments. The analysis identifies information preserved by whole-group corruption and limits on interpreting our asymmetric sweep as a precision-recall comparison. Overall, the results indicate that imperfect verification can support effective RLVR in the tested settings, while aggregate error rates alone do not characterize the learning signal.

cs.LG↗

arXiv2026

A Cheap Verifier is Good Enough: LLM Post-training is Robust to Erroneous Rewards

When post-training large language models on tasks with semi-verifiable rewards, there are many factors (training steps, base model size, training order, data quality, verifier accuracy, etc.) that practitioners must contend with to maximize model performance. Yet, it remains unclear how well verifier agreement predicts post-training performance on such tasks. In this paper, we explore this question with over 11k H100 GPU-hours, across HealthBench and PRBench tasks in medical, legal, and finance domains. Across the tested domains, Qwen3 trainees (1.7B-8B on HealthBench; 8B on PRBench), evaluation splits, and frontier LLM reference judges (which we call golden verifiers), higher verifier agreement does not consistently identify the best training verifier. Expensive verifiers need not outperform inexpensive ones, and open-weight Gemma verifiers produce strong training outcomes. We compare two low-cost choices retrospectively -- a cost-reducing choice and a balanced choice -- with estimated grading cost reductions of 98.8%-99.7% relative to the golden grading protocols and average post-training score gaps of 1-3 points from the best evaluated training verifier. These averages include larger losses in individual settings; they do not establish that verifier choices are interchangeable.

cs.LG↗

arXiv2026

VIALS: A Benchmark for Visual Interpretation of Artifacts in the Life Sciences

In professional life sciences workflows, scientists routinely interpret visual artifacts (gel blots, microscopy images, plasmid maps, flow cytometry plots, molecular structures, ...) to inform research decisions. We introduce VIALS, a visual question-answering benchmark with 161 such interpretation tasks, spanning the types of artifacts examined throughout experimental workflows in the biotech industry (rather than polished figures from publications and textbooks). While frontier vision-language models can now fluently describe natural images, we find that they are unable to accurately interpret these scientific images, reflecting limitations in domain knowledge and domain-specific visual reasoning capabilities. In contrast, scientists with relevant domain expertise find these visual interpretation tasks straightforward. AI that cannot similarly interpret such images will have limited utility in professional life sciences workflows, where such artifacts are central to how scientists reason, communicate, and make decisions.

cs.AI↗

arXiv2026

CAREBench: A Child-Safety Risk Benchmark for Language Models

How can we evaluate whether frontier AI systems recognize child-safety risks before they escalate into explicit harm? Existing child safety evaluations focus on child sexual abuse material, yet many child-safety failures begin earlier: in model assistance that helps adults manipulate, impersonate, profile, or isolate minors, and in model responses that deepen children's emotional dependence on AI systems rather than redirecting them toward human support. We introduce CAREBench (Child AI Risk Evaluation), a benchmark to assess such upstream child-safety risks in language models. CAREBench contains 500 prompts spanning twelve risk categories, including grooming and relationship engineering, deception and impersonation, surveillance and privacy, sextortion and sexual abuse, AI anthropomorphization, emotional dependency, and mental illness sensitivity. Developed with response annotations from parents and clinicians, the benchmark excludes explicit abuse material and imagery; instead, it evaluates whether models recognize, refuse, de-escalate, or redirect risky interactions before harm becomes overt. Evaluating seven frontier models on our benchmark, we find failure rates ranging from 2% to 58%, with failure patterns that vary across risk categories. CAREBench provides a responsibly scoped evaluation for LLM developers to identify and close gaps in child safety policies.

cs.LG↗

arXiv2026

BankerToolBench: Evaluating AI Agents in End-to-End Investment Banking Workflows

Existing AI benchmarks lack the fidelity to assess economically meaningful progress on professional workflows. To evaluate frontier AI agents in a high-value, labor-intensive profession, we introduce BankerToolBench (BTB): an open-source benchmark of end-to-end analytical workflows routinely performed by junior investment bankers. To develop an ecologically valid benchmark grounded in representative work environments, we collaborated with 502 investment bankers from leading firms. BTB requires agents to execute senior banker requests by navigating data rooms, using industry tools (market data platform, SEC filings database), and generating multi-file deliverables--including Excel financial models, PowerPoint pitch decks, and PDF/Word reports. Completing a BTB task takes bankers up to 21 hours, underscoring the economic stakes of successfully delegating this work to AI. BTB enables automated evaluation of any LLM or agent, scoring deliverables against 100+ rubric criteria defined by veteran investment bankers to capture stakeholder utility. Testing 9 frontier models, we find that even the best-performing model (GPT-5.4) fails nearly half of the rubric criteria and bankers rate 0% of its outputs as client-ready. Our failure analysis reveals key obstacles (such as breakdowns in cross-artifact consistency) and improvement directions for agentic AI in high-stakes professional workflows.

cs.AI↗

arXiv2026

MENLO: From Preferences to Proficiency -- Evaluating and Modeling Native-like Quality Across 47 Languages

Ensuring native-like quality of large language model (LLM) responses across many languages is challenging. To address this, we introduce MENLO, a framework that operationalizes the evaluation of native-like response quality based on audience design-inspired mechanisms. Using MENLO, we create a dataset of 6,423 human-annotated prompt-response preference pairs covering four quality dimensions with high inter-annotator agreement in 47 language varieties. Our evaluation reveals that zero-shot LLM judges benefit significantly from pairwise evaluation and our structured annotation rubrics, yet they still underperform human annotators on our dataset. We demonstrate substantial improvements through fine-tuning with reinforcement learning, reward shaping, and multi-task learning approaches. Additionally, we show that RL-trained judges can serve as generative reward models to enhance LLMs' multilingual proficiency, though discrepancies with human judgment remain. Our findings suggest promising directions for scalable multilingual evaluation and preference alignment. We release our dataset and evaluation framework to support further research in multilingual LLM evaluation (https://huggingface.co/datasets/facebook/menlo).

cs.CL↗

arXiv2026

Translation as a Scalable Proxy for Multilingual Evaluation

The rapid proliferation of LLMs has created a critical evaluation paradox: while LLMs claim multilingual proficiency, comprehensive non-machine-translated benchmarks exist for fewer than 30 languages, leaving >98% of the world's 7,000 languages in an empirical void. Traditional benchmark construction faces scaling challenges such as cost, scarcity of domain experts, and data contamination. We evaluate the validity of a simpler alternative: can translation quality alone indicate a model's broader multilingual capabilities? Through systematic evaluation of 14 models (1B-72B parameters) across 9 diverse benchmarks and 7 translation metrics, we find that translation performance is a good indicator of downstream task success (e.g., Phi-4, median Pearson r: MetricX = 0.89, xCOMET = 0.91, SSA-COMET = 0.87). These results suggest that the representational abilities supporting faithful translation overlap with those required for multilingual understanding. Translation quality, thus emerges as a strong, inexpensive first-pass proxy of multilingual performance, enabling a translation-first screening with targeted follow-up for specific tasks.

cs.CL↗

arXiv2024

The Llama 3 Herd of Models

Modern artificial intelligence (AI) systems are powered by foundation models. This paper presents a new set of foundation models, called Llama 3. It is a herd of language models that natively support multilinguality, coding, reasoning, and tool usage. Our largest model is a dense Transformer with 405B parameters and a context window of up to 128K tokens. This paper presents an extensive empirical evaluation of Llama 3. We find that Llama 3 delivers comparable quality to leading language models such as GPT-4 on a plethora of tasks. We publicly release Llama 3, including pre-trained and post-trained versions of the 405B parameter language model and our Llama Guard 3 model for input and output safety. The paper also presents the results of experiments in which we integrate image, video, and speech capabilities into Llama 3 via a compositional approach. We observe this approach performs competitively with the state-of-the-art on image, video, and speech recognition tasks. The resulting models are not yet being broadly released as they are still under development.

cs.AI↗

arXiv2023

Seamless: Multilingual Expressive and Streaming Speech Translation

Large-scale automatic speech translation systems today lack key features that help machine-mediated communication feel seamless when compared to human-to-human dialogue. In this work, we introduce a family of models that enable end-to-end expressive and multilingual translations in a streaming fashion. First, we contribute an improved version of the massively multilingual and multimodal SeamlessM4T model-SeamlessM4T v2. This newer model, incorporating an updated UnitY2 framework, was trained on more low-resource language data. SeamlessM4T v2 provides the foundation on which our next two models are initiated. SeamlessExpressive enables translation that preserves vocal styles and prosody. Compared to previous efforts in expressive speech research, our work addresses certain underexplored aspects of prosody, such as speech rate and pauses, while also preserving the style of one's voice. As for SeamlessStreaming, our model leverages the Efficient Monotonic Multihead Attention mechanism to generate low-latency target translations without waiting for complete source utterances. As the first of its kind, SeamlessStreaming enables simultaneous speech-to-speech/text translation for multiple source and target languages. To ensure that our models can be used safely and responsibly, we implemented the first known red-teaming effort for multimodal machine translation, a system for the detection and mitigation of added toxicity, a systematic evaluation of gender bias, and an inaudible localized watermarking mechanism designed to dampen the impact of deepfakes. Consequently, we bring major components from SeamlessExpressive and SeamlessStreaming together to form Seamless, the first publicly available system that unlocks expressive cross-lingual communication in real-time. The contributions to this work are publicly released and accessible at https://github.com/facebookresearch/seamless_communication

cs.CL↗

arXiv2023

SeamlessM4T: Massively Multilingual & Multimodal Machine Translation

What does it take to create the Babel Fish, a tool that can help individuals translate speech between any two languages? While recent breakthroughs in text-based models have pushed machine translation coverage beyond 200 languages, unified speech-to-speech translation models have yet to achieve similar strides. More specifically, conventional speech-to-speech translation systems rely on cascaded systems that perform translation progressively, putting high-performing unified systems out of reach. To address these gaps, we introduce SeamlessM4T, a single model that supports speech-to-speech translation, speech-to-text translation, text-to-speech translation, text-to-text translation, and automatic speech recognition for up to 100 languages. To build this, we used 1 million hours of open speech audio data to learn self-supervised speech representations with w2v-BERT 2.0. Subsequently, we created a multimodal corpus of automatically aligned speech translations. Filtered and combined with human-labeled and pseudo-labeled data, we developed the first multilingual system capable of translating from and into English for both speech and text. On FLEURS, SeamlessM4T sets a new standard for translations into multiple target languages, achieving an improvement of 20% BLEU over the previous SOTA in direct speech-to-text translation. Compared to strong cascaded models, SeamlessM4T improves the quality of into-English translation by 1.3 BLEU points in speech-to-text and by 2.6 ASR-BLEU points in speech-to-speech. Tested for robustness, our system performs better against background noises and speaker variations in speech-to-text tasks compared to the current SOTA model. Critically, we evaluated SeamlessM4T on gender bias and added toxicity to assess translation safety. Finally, all contributions in this work are open-sourced and accessible at https://github.com/facebookresearch/seamless_communication

cs.CL↗

arXiv2023

The 2023 release of Cloudy

We describe the 2023 release of the spectral synthesis code Cloudy. Since the previous major release, migrations of our online services motivated us to adopt git as our version control system. This change alone led us to adopt an annual release scheme, accompanied by a short release paper, the present being the inaugural. Significant changes to our atomic and molecular data have improved the accuracy of Cloudy predictions: we have upgraded our instance of the Chianti database from version 7 to 10; our H- and He-like collisional rates to improved theoretical values; our molecular data to the most recent LAMDA database, and several chemical reaction rates to their most recent UDfA and KiDA values. Finally, we describe our progress on upgrading Cloudy's capabilities to meet the requirements of the X-ray microcalorimeters aboard the upcoming XRISM and Athena missions, and outline future development that will make Cloudy of use to the X-ray community.

astro-ph.GA↗

arXiv2022

No Language Left Behind: Scaling Human-Centered Machine Translation

Driven by the goal of eradicating language barriers on a global scale, machine translation has solidified itself as a key focus of artificial intelligence research today. However, such efforts have coalesced around a small subset of languages, leaving behind the vast majority of mostly low-resource languages. What does it take to break the 200 language barrier while ensuring safe, high quality results, all while keeping ethical considerations in mind? In No Language Left Behind, we took on this challenge by first contextualizing the need for low-resource language translation support through exploratory interviews with native speakers. Then, we created datasets and models aimed at narrowing the performance gap between low and high-resource languages. More specifically, we developed a conditional compute model based on Sparsely Gated Mixture of Experts that is trained on data obtained with novel and effective data mining techniques tailored for low-resource languages. We propose multiple architectural and training improvements to counteract overfitting while training on thousands of tasks. Critically, we evaluated the performance of over 40,000 different translation directions using a human-translated benchmark, Flores-200, and combined human evaluation with a novel toxicity benchmark covering all languages in Flores-200 to assess translation safety. Our model achieves an improvement of 44% BLEU relative to the previous state-of-the-art, laying important groundwork towards realizing a universal translation system. Finally, we open source all contributions described in this work, accessible at https://github.com/facebookresearch/fairseq/tree/nllb.

cs.CL↗

arXiv2022

X-ray spectroscopy in the microcalorimeter era 4: Optical depth effects on the soft X-rays studied with CLOUDY

In this paper, we discuss atomic processes modifying the soft X-ray spectra in from optical depth effects like photoelectric absorption and electron scattering suppressing the soft X-ray lines. We also show the enhancement in soft X-ray line intensities in a photoionized environment via continuum pumping. We quantify the suppression/enhancement by introducing a "line modification factor ($f_{\rm mod}$)." If 0 $\leq$ $f_{\rm mod}$ $\leq$ 1, the line is suppressed, which could be the case in both collisionally-ionized and photoionized systems. If $f_{\rm mod}$ $\geq$ 1, the line is enhanced, which occurs in photoionized systems. Hybrid astrophysical sources are also very common, where the environment is partly photoionized and partly collisionally-ionized. Such a system is V1223 Sgr, an intermediate polar binary. We show the application of our theory by fitting the first-order Chandra MEG spectrum of V1223 Sgr with a combination of \textsc{Cloudy}-simulated additive cooling-flow and photoionized models. In particular, we account for the excess flux for O~VII, O~VIII, Ne~IX, Ne~X, and Mg~XI lines in the spectrum found in a recent study, which could not be explained with an absorbed cooling-flow model.

astro-ph.HE↗

arXiv2022

OCR Improves Machine Translation for Low-Resource Languages

We aim to investigate the performance of current OCR systems on low resource languages and low resource scripts. We introduce and make publicly available a novel benchmark, OCR4MT, consisting of real and synthetic data, enriched with noise, for 60 low-resource languages in low resource scripts. We evaluate state-of-the-art OCR systems on our benchmark and analyse most common errors. We show that OCR monolingual data is a valuable resource that can increase performance of Machine Translation models, when used in backtranslation. We then perform an ablation study to investigate how OCR errors impact Machine Translation performance and determine what is the minimum level of OCR quality needed for the monolingual data to be useful for Machine Translation.

cs.CL↗

arXiv2021

MLQE-PE: A Multilingual Quality Estimation and Post-Editing Dataset

We present MLQE-PE, a new dataset for Machine Translation (MT) Quality Estimation (QE) and Automatic Post-Editing (APE). The dataset contains eleven language pairs, with human labels for up to 10,000 translations per language pair in the following formats: sentence-level direct assessments and post-editing effort, and word-level good/bad labels. It also contains the post-edited sentences, as well as titles of the articles where the sentences were extracted from, and the neural MT models used to translate the text.

cs.CL↗

arXiv2021

Classification-based Quality Estimation: Small and Efficient Models for Real-world Applications

Sentence-level Quality estimation (QE) of machine translation is traditionally formulated as a regression task, and the performance of QE models is typically measured by Pearson correlation with human labels. Recent QE models have achieved previously-unseen levels of correlation with human judgments, but they rely on large multilingual contextualized language models that are computationally expensive and make them infeasible for real-world applications. In this work, we evaluate several model compression techniques for QE and find that, despite their popularity in other NLP tasks, they lead to poor performance in this regression setting. We observe that a full model parameterization is required to achieve SoTA results in a regression task. However, we argue that the level of expressiveness of a model in a continuous range is unnecessary given the downstream applications of QE, and show that reframing QE as a classification problem and evaluating QE models using classification metrics would better reflect their actual performance in real-world applications.

cs.CL↗

arXiv2021

XLEnt: Mining a Large Cross-lingual Entity Dataset with Lexical-Semantic-Phonetic Word Alignment

Cross-lingual named-entity lexica are an important resource to multilingual NLP tasks such as machine translation and cross-lingual wikification. While knowledge bases contain a large number of entities in high-resource languages such as English and French, corresponding entities for lower-resource languages are often missing. To address this, we propose Lexical-Semantic-Phonetic Align (LSP-Align), a technique to automatically mine cross-lingual entity lexica from mined web data. We demonstrate LSP-Align outperforms baselines at extracting cross-lingual entity pairs and mine 164 million entity pairs from 120 different languages aligned with English. We release these cross-lingual entity pairs along with the massively multilingual tagged named entity corpus as a resource to the NLP community.

cs.CL↗