Fly-Swat or Cannon? Cost-Effective Language Model Choice via Meta-Modeling

Chat with Graph Search

Ask any question about EPFL courses, lectures, exercises, research, news, etc. or try the example questions below.

DISCLAIMER: The Graph Chatbot is not programmed to provide explicit or categorical answers to your questions. Rather, it transforms your questions into API requests that are distributed across the various IT services officially administered by EPFL. Its purpose is solely to collect and recommend relevant references to content that you can explore to help you answer your questions.

We study the extent to which credit index (CDX) options are priced consistent with S&P 500 (SPX) equity index options. We derive analytical expressions for CDX and SPX options within a structural credit-risk model with stochastic volatility and jumps using ...

The ability to reason, plan and solve highly abstract problems is a hallmark of human intelligence. Recent advancements in artificial intelligence, propelled by deep neural networks, have revolutionized disciplines like computer vision and natural language ...

Accurate human trajectory prediction is crucial for applications such as autonomous vehicles, robotics, and surveillance systems. Yet, existing models often fail to fully leverage the non-verbal social cues human subconsciously communicate when navigating ...

Transformer models such as GPT generate human-like language and are predictive of human brain responses to language. Here, using functional-MRI-measured brain responses to 1,000 diverse sentences, we first show that a GPT-based encoding model can predict t ...

The production of aluminium largely depends on the use of fossil fuels, resulting in the emission of significant amounts of greenhouse gases. As the aluminium industry is working towards decreasing its environmental burdens, the elimination of direct emiss ...

The scale and pervasiveness of the Internet make it a pillar of planetary communication, industry and economy, as well as a fundamental medium for public discourse and democratic engagement. In stark contrast with the Internet's decentralized infrastructur ...

Cell segmentation is a critical step for quantitative single-cell analysis in microscopy images. Existing cell segmentation methods are often tailored to specific modalities or require manual interventions to specify hyper-parameters in different experimen ...

Natural language processing and other artificial intelligence fields have witnessed impressive progress over the past decade. Although some of this progress is due to algorithmic advances in deep learning, the majority has arguably been enabled by scaling ...

In this work we consider an online learning problem, called Online k-Clustering with Moving Costs, at which a learner maintains a set of k facilities over T rounds so as to minimize the connection cost of an adversarially selected sequence of clients. The ...

Time has always been a central factor in understanding the challenges of daily mobility. For a long time, and still today, methods of economic evaluation of transport projects have monetized time savings so that they can be included in the cost–benefit ana ...

A strategic roadmap for noncarbonized fuels is a global priority, and the reduction of carbon dioxide emissions is a key focus of the Paris Agreement to mitigate the effects of rising temperatures. In this context, hydrogen is a promising noncarbonized fue ...

Recent transformer language models achieve outstanding results in many natural language processing (NLP) tasks. However, their enormous size often makes them impractical on memory-constrained devices, requiring practitioners to compress them to smaller net ...

In this dissertation, we propose multiple methods to improve transfer learning for pretrained language models (PLMs). Broadly, transfer learning is a powerful technique in natural language processing, where a language model is first pre-trained on a data-r ...

Natural language processing has experienced significant improvements with the development of Transformer-based models, which employ self-attention mechanism and pre-training strategies. However, these models still present several obstacles. A notable issue ...

In this article, we account for the liquidity risk in the underlying assets when pricing European exchange options, which has not been considered in the literature. An Ornstein-Uhlenbeck process with the mean -reversion property is selected to model the ma ...

The sequence of codes Serpent/DYN3D has been developed by the Helmholtz-Zentrum Dresden-Rossendorf and successfully applied to core static and transient analyses of sodium-cooled fast reactors (SFRs). The successful application of the sequence to SFRs was ...

Abstractive summarization has seen big improvements in recent years, mostly due to advances in neural language modeling, language model pretraining, and scaling models and datasets. While large language models generate summaries that are fluent, coherent, ...

This dataset complements the article "Frequency regulation with storage: On losses and profits" by Dirk Lauinger, François Vuille, and Daniel Kuhn, available at https://arxiv.org/abs/2306.02987. The dataset contains the following files: 1.

K-means is one of the fundamental unsupervised data clustering and machine learning methods. It has been well studied over the years: parallelized, approximated, and optimized for different cases and applications. With increasingly higher parallelism leadi ...

In this thesis we present three closed form approximation methods for portfolio valuation and risk management. The first chapter is titled ``Kernel methods for portfolio valuation and risk management'', and is a joint work with Damir Filipovi'c (SFI and ...

Fly-Swat or Cannon? Cost-Effective Language Model Choice via Meta-Modeling

Chat with Graph Search

How Integrated are Credit and Equity Markets? Evidence from Index Options

Infusing structured knowledge priors in neural models for sample-efficient symbolic reasoning

Social-Transmotion: Promptable Human Trajectory Prediction

Driving and suppressing the human language network using large language models

CO2 Capture and Management Strategies for Decarbonizing Secondary Aluminium Production

Planetary-Scale Byzantine Fault Tolerance

The multimodality cell segmentation challenge: toward universal solutions

Text Representation Learning for Low Cost Natural Language Understanding

Efficient Online Clustering with Moving Costs

Appropriations and Uses of Travel Time

Updates on Hydrogen Value Chain: A Strategic Roadmap

Revisiting Offline Compression: Going Beyond Factorization-based Methods for Transformer Language Models

Improving Generalization of Pretrained Language Models

Modeling Structured Data in Attention-based Models

Exchange options with stochastic liquidity risk

Extension, validation, and optimization of Serpent/DYN3D/ATHLET code system for SFR applications

Interpretable Representation Learning and Evaluation for Abstractive Summarization

Dataset for the article "Frequency regulation with storage: On losses and profits"

Improving K-means Clustering Using Speculation

Closed form approximation methods for portfolio valuation and risk management