Research
On-device research index

arXiv research

A locally-built, LLM-digested index of recent arXiv papers in quant finance, geometry/topology, and statistical ML — keyword search served straight from SQLite on this machine.

168,657 papers · 148 categories

Trend · papers per month

4228441,2651,687 · Jun 202019922001200920172026
48 results for Language model training

Optimus pre-trains sentences in a latent space for various NLP tasks.

problem Training large-scale language models for diverse NLP tasks.
method Pre-trained Variational Autoencoder (VAE) on large text corpus, fine-tuned for various tasks.
result Optimus achieves state-of-the-art on VAE language modeling benchmarks.

Chronologically consistent models maintain accuracy with time-restricted data.

problem Training data introduces lookahead bias and training leakage in large language models.
method ChronoBERT and ChronoGPT trained with only available data at each time point.
result Models achieve strong performance and competitive with larger models, mitigating lookahead bias.

Self-supervised skip-tree training improves mathematical reasoning in language models.

problem Improving logical reasoning in language models for formal mathematics.
method Self-supervised language modeling on mathematical formulas, skip-tree task.
result Models trained on skip-tree task outperform standard models in mathematical reasoning tasks.

Generative Adversarial Networks (GANs) enjoy great success at image generation, but have proven difficult to train in the domain of natural language. Challenges with gradient estimation, optimization instability, and mode collapse have lead practitioners to resort to maximum likelihood pre-training, followed by small a…

2019-05-23abs ↗pdf ↗

Improved ASR for English-isiZulu code-switched speech with semi-supervised training.

problem Improving ASR for code-switched speech between English and isiZulu.
method Semi-supervised training using automatic transcription of multilingual speech data.
result Semi-supervised training achieved significant WER reduction in ASR performance.

Recently, pre-trained language representation flourishes as the mainstay of the natural language understanding community, e.g., BERT. These pre-trained language representations can create state-of-the-art results on a wide range of downstream tasks. Along with continuous significant performance improvement, the size an…

2019-09-27abs ↗pdf ↗

New findings show language models can't simultaneously avoid hallucinations and capture all language richness.

problem Achieving both valid output and full language richness in language generation.
method Investigates language generation within a statistical setting, focusing on consistency and breadth.
result For most collections of candidate languages, a language model cannot simultaneously avoid hallucinations and capture all language richness.

Prototype model improves model auditing and understanding.

problem Auditing and understanding modern language models is expensive and approximate.
method Introduced a sparse, non-negative mixture of learned prototypes trained with clustering objectives.
result Prototype models either surpass or remain within 2.5 percentage points of dense baselines on downstream tasks.

Work shows hallucination detection by LLMs is impossible without expert feedback.

problem Detecting hallucinations in LLMs is theoretically impossible without expert-labeled feedback.
method Investigated hallucination detection using a theoretical framework inspired by language identification.
result Automated hallucination detection is impossible for most language collections without expert-labeled feedback.

When a bilingual student learns to solve word problems in math, we expect the student to be able to solve these problem in both languages the student is fluent in,even if the math lessons were only taught in one language. However, current representations in machine learning are language dependent. In this work, we pres…

2018-09-23abs ↗pdf ↗

New attack reveals memorization patterns in pre-trained LLMs.

problem Determining if a data point was part of a pre-trained LLM's training set.
method Adapts MIA statistical tests to LLM's perplexity dynamics of subsequences.
result Significantly outperforms prior approaches in membership inference attacks.

Improved speech recognition with language model integration in sequence-to-sequence models.

problem Improving word error rate in speech recognition models.
method Log-linear combination of acoustic and language models with per-token renormalization.
result The proposed method shows good improvements over standard model combination on Librispeech system.

xVal tokenizes numbers continuously for better scientific model training.

problem Lack of continuous numerical tokenization for scientific datasets in LLMs.
method xVal: Continuous numerical tokenization strategy.
result xVal outperforms other numerical tokenization methods on scientific datasets.

ScheduleFree+ improves large language model training without schedules or learning rates.

problem Scaling up Schedule-Free Learning to large language models.
method Learning-rate-free and schedule-free method for training large language models.
result ScheduleFree+ outperforms SOTA schedules by 31% at 1000 tokens per parameter.

Adaptive batch size schedules improve language model training efficiency and generalization.

problem Dilemma of choosing batch sizes in large-scale model training.
method General-purpose adaptive batch size schedules compatible with data and model parallelism.
result Adaptive batch size schedules outperform constant batch sizes and heuristic warmup schedules.

BloombergGPT is a large language model trained on financial data, outperforming existing models on financial tasks.

problem Lack of specialized large language models for finance.
method Trained on a 363 billion token dataset augmented with 345 billion tokens from general datasets, using a 50 billion parameter model.
result BloombergGPT outperforms existing models on financial tasks without sacrificing performance on general LLM benchmarks.

Unlike major Western languages, most African languages are very low-resourced. Furthermore, the resources that do exist are often scattered and difficult to obtain and discover. As a result, the data and code for existing research has rarely been shared. This has lead a struggle to reproduce reported results, and few p…

2019-06-17abs ↗pdf ↗

Large language models have recently achieved state of the art performance across a wide variety of natural language tasks. Meanwhile, the size of these models and their latency have significantly increased, which makes their usage costly, and raises an interesting question: do language models need to be large? We study…

2019-10-10abs ↗pdf ↗

New study finds best language model architecture and pretraining objective for zero-shot tasks.

problem Evaluating which language model architectures and pretraining objectives best enable zero-shot generalization.
method Compared three model architectures and two pretraining objectives across 170 billion tokens, with and without finetuning.
result Causal decoder-only models trained on autoregressive language modeling exhibit strongest zero-shot generalization.

GNMR controls runtime stability in low-precision language model training.

problem Efficient low-precision training faces numerical risks at specific operators.
method GNMR compares gradient norms to historical means, applying bounded recovery actions.
result GNMR preserves high-fidelity quality with sparse, budgeted recovery.

Information extraction is an important task in NLP, enabling the automatic extraction of data for relational database filling. Historically, research and data was produced for English text, followed in subsequent years by datasets in Arabic, Chinese (ACE/OntoNotes), Dutch, Spanish, German (CoNLL evaluations), and many …

2019-11-19abs ↗pdf ↗

Enhances drug discovery models by understanding human language.

problem Low predictive quality of activity prediction models in drug discovery.
method Proposes a novel architecture with separate chemical and natural language input modules and a contrastive pre-training objective.
result Improves predictive performance on few-shot and zero-shot learning benchmarks.

Fine-tuning large language models requires minimal data, making them efficient.

problem Achieving state-of-the-art performance with large language models.
method Using BERT as an example, fine-tuning only the most critical layers of the pre-trained model.
result Fine-tuned models are close in parameter space to the pre-trained model, with many good solutions found in sparsified versions.

Paper reviews neurolinguistics and language technologies, emphasizing mutual enrichment.

problem Understanding brain activity during language processing.
method Brain imaging studies and natural language representations.
result Development of brain-aware natural language representations.

Researchers show how to secretly train models with hidden data, detect usage with high confidence.

problem Protecting training data from traceability in large language models.
method Gradient-based optimization to learn secret sequences absent from training data.
result Secret sequences can be learned by models without performance degradation, detectable with high confidence.

RecoBERT uses a language model to recommend items from catalogs.

problem Harnessing language models for text-based item recommendations.
method RecoBERT is a BERT-based approach that learns specialized language models for item recommendations without requiring labeled data.
result RecoBERT outperforms other techniques in inferring item similarities from textual catalogs.

Language models trained on chess board states outperform those on moves, even with causal masking.

problem Applying causal masking to spatial data for training unimodal language models.
method Trained bidirectional and causal self-attention models on both spatial (board-based) and sequential (move-based) chess data.
result Models trained on spatial board states achieve stronger playing strength than those trained on sequential data, even with causal masking.

The study examines how language models learn to represent the world, identifying conditions for ecological veridicality.

problem Understanding when language models learn to represent the world accurately and how this learning process can fail.
method Analyzes the Bayes-optimal next-token cross-entropy decomposition and the role of training ecology in shaping model representations.
result The minimum-complexity zero-excess solution is the quotient partition by training equivalence, and this solution is not preserved in in-context learning or per-task adaptation.

Recently, substantial progress has been made in language modeling by using deep neural networks. However, in practice, large scale neural language models have been shown to be prone to overfitting. In this paper, we present a simple yet highly effective adversarial training mechanism for regularizing neural language mo…

2019-06-10abs ↗pdf ↗