我有一个 PPT（beamer 类），我想向其中添加 bib 文件 ( main.bib)。运行 BibTeX 时出现以下错误
I found no \citation commands---while reading file masterthesis.aux
根据 tex 网站的建议，我\nocite{*}在 \bibliography 命令之前添加了它。我的 tex 代码的这一部分如下所示
\bibliographystyle{unsrt}
\nocite{*}
\bibliography{main}
在代码顶部，包含的包如下：-
\documentclass[11pt]{beamer}
\usetheme{Berkeley}
\usepackage[utf8]{inputenc}
\usepackage[english]{babel}
\usepackage{amsmath}
\usepackage{amsfonts}
\usepackage{datetime}
\usepackage{amssymb}
\usepackage{graphicx}
\usepackage{ragged2e}
该\bibliography命令对我来说看起来很可疑，因为它在 TexMaker 控制台中以不同的颜色出现。
有什么想法吗？我也附上了代码
\documentclass[11pt]{beamer}
\usetheme{Berkeley}
\usepackage[utf8]{inputenc}
\usepackage[english]{babel}
\usepackage{amsmath}
\usepackage{amsfonts}
\usepackage{datetime}
\usepackage{amssymb}
\usepackage{graphicx}
\usepackage{ragged2e}
\author{Upendra Pratap Singh}
\title{Transfer Learning}
%\setbeamertemplate{navigation symbols}{\insertlogo}
%\setbeamercovered{transparent}
%\setbeamertemplate{navigation symbols}{}
\logo{\includegraphics[height=1.7cm]{logo.png}}
\institute{Indian Institute of Information Technology, Allahabad}
\date{\today}
\subject{Data Analytics and Machine Learning}
\begin{document}
\begin{frame}
\titlepage
\end{frame}
%
%\begin{frame}
%\tableofcontents
%\end{frame}
\begin{frame}{Introduction}
\begin{figure}[h!]
\centering
\includegraphics[scale=0.4]{andrew_nips.png}
\caption{sample}
\end{figure}
\end{frame}
\begin{frame}{Transfer Learning - Psychological Point of View}
\begin{enumerate}
\item The study of dependency of human conduct, learning or performance on prior experience
\item Describes the effect of past learning upon present acquisition
\item Can be partial/complete application or carryover of knowledge/skills
\end{enumerate}
\end{frame}
\begin{frame}{Transfer Learning versus Traditional Machine Learning paradigm}
\begin{figure}[h!]
\centering
\includegraphics[scale=0.5]{traditional_ml.png}
\caption{Traditional Machine Learning Paradigm}
\end{figure}
\end{frame}
\begin{frame}{Transfer Learning versus Traditional Machine Learning paradigm}
\begin{figure}[h!]
\centering
\includegraphics[scale=0.5]{transfer_learning_paradigm.png}
\caption{Transfer Learning Paradigm}
\end{figure}
\end{frame}
\begin{frame}{Transfer Learning - Psychological Point of View}
Types of Transfer Learning
\begin{enumerate}
\item Positive transfer
\item Negative transfer
\item Neutral transfer
\end{enumerate}
\end{frame}
\begin{frame}{Transfer Learning - Psychological Point of View}
Theories of Transfer Learning
\begin{enumerate}
\item Theory of identical elements
\begin{itemize}
\item By E.L. Thorndike
\item Carrying over from one situation to another is roughly proportional to the degree of resemblance of a situation
\end{itemize}
\end{enumerate}
\end{frame}
\begin{frame}{Transfer Learning - Psychological Point of View}
Theories of Transfer Learning
\begin{enumerate}
\item Theory of generalization of experience
\begin{itemize}
\item By Charles Judd
\item Assumes that what is learnt in task A transfers to task B, because in studying task A, the learner develops a general principle which applies in part or completely in both A and B
\item The ability of individuals to generalize knowledge varies with the degree of their intelligence
\end{itemize}
\end{enumerate}
\end{frame}
\begin{frame}{Transfer Learning - Machine Learning perspective}
\begin{figure}[h!]
\centering
\includegraphics[scale=0.5]{transfer_of_learning.png}
\caption{Transfer Learing - Machine Learning Perspective}
\end{figure}
The ability of a system to recognize and apply knowledge and skills learned in previous domains/tasks to novel tasks/domains which share some commonality.
\end{frame}
\begin{frame}{Transfer Learning - A two-step process}
\begin{enumerate}
\item To identify the commonality between the tasks/domains
\item To transfer knowledge from previous domains to the target domains
\end{enumerate}
\end{frame}
\begin{frame}{Why Transfer Learning}
\begin{enumerate}
\item Datasets unavailable or expensive to procure
\item Theoretical test data is different from real world test data
\item Real world -messy
\item Each client has individual preferences - thus test data unique
\end{enumerate}
As a result, models frightfully lack the ability to generalize to real world scenarios
\end{frame}
\begin{frame}{Evolution of Transfer Learning}
\begin{enumerate}
\item Tom Mitchell proposed \textit{bias learning}
\item Neural Nets used for transfer learning
\item Pratt adopted \textit{entropy} measures to assess the quality of hyperplanes
\item Framework developed to use internal representations generated by NNs for tranfer learning problems
\end{enumerate}
\end{frame}
\begin{frame}{Evolution of Transfer Learning}
\begin{enumerate}
\item \textit{Incremental learning} concepts were formulated that result in extreme nested system of learning functions
\item \textit{Multi-Task Learning} proposed
\item \textit{Covariate shift} coined by Shimodiara
\item \textit{Domain Adaptation} evolved to get rid of limitations of covariate shift
\end{enumerate}
\end{frame}
\begin{frame}{Evolution of Transfer Learning}
\begin{enumerate}
\item \textit{Leave One Out Risk} was overcome
\item Domain adaptation applied intensively to NLP problems
\end{enumerate}
\end{frame}
\begin{frame}{Transfer Learning Strategies}
\begin{figure}[h!]
\centering
\includegraphics[scale=0.5]{transfer_learning_strategies.png}
\caption{Tranfer Learning Strategies in a Nutshell}
\end{figure}
\end{frame}
\begin{frame}{Transfer Learning Strategies}
\begin{enumerate}
\item To correct for marginal distribution difference in the source
\item To correct for conditional distribution difference in the source
\item To correct both the marginal and conditional distribution in the source
\end{enumerate}
\end{frame}
\begin{frame}{Information Transfer - What is being transferred}
Instances
\begin{enumerate}
\item Instance are re-weighted to correct for marginal distribution differences
\item The reweighted instances are directly used in the target domain
\end{enumerate}
\end{frame}
\begin{frame}{Information Transfer - What is being transferred}
Features - First Approach
\begin{enumerate}
\item Transform the features of source
\item Reweight the features so that they more closely match the target domain
\item Above process also referred to as \textit{asymmetric feature transformation}
\end{enumerate}
\end{frame}
\begin{frame}{Information Transfer - What is being transferred}
Features - Second Approach
\begin{enumerate}
\item Discover underlying meaningful structures between domains to find a common latent feature space that has predictive qualities
\item Reduce the marginal distribution between the domains at the same time
\item Above approach - \textit{symmetric feature transformation}
\end{enumerate}
\end{frame}
\begin{frame}{Information Transfer - What is being transferred}
Parameters
\begin{enumerate}
\item Transfer knowledge through shared parameters of source and target domain learner models or
\item Create multiple source learner models and optimally combine the reweighted learners to improve the target learner
\end{enumerate}
\end{frame}
\begin{frame}{Neural Networks and Transfer Learning}
\begin{enumerate}
\item NNs are the majority algorithms to perform Transfer learning.
\item Key aspects to be considered are as follows
\begin{itemize}
\item How can we tailor deep neural networks for transfer learning ?
\item How transfer learning performs with reusing layers and with different types of data
\end{itemize}
\end{enumerate}
\end{frame}
\begin{frame}{Neural Networks and Transfer Learning}
\begin{enumerate}
\item NNs gain knowledge from the data it is trained on
\item Knowledge gets compiled as weights
\item Weights can be extracted and then transferred to any other neural network
\item So, no need to train NN from scratch
\end{enumerate}
\end{frame}
\begin{frame}{Neural Networks and Transfer Learning}
Ways to fine tune the Neural Network model
\begin{enumerate}
\item \textit{Feature Extractor: } remove output layer of neural network and then use entire network as a fixed feature extractor
\item Use \textit{entire} architecture, \textit{initialize weight randomly} and train the model
\item Retrain some layers while freeze others: keep weights of initial layers as constant and modify only the higher layers
\end{enumerate}
\end{frame}
\begin{frame}{Trends and Challenges}
\begin{enumerate}
\item \textit{Knowledge gain} when doing transfer learning and its quantification
\item \textit{Impact} of different datasets (target problems) on learning rates
\item \textit{Benchmark} available methods
\item \textit{Unification} of Transfer Learning due to vast amount of formulations available
\item Public competition for \textit{benchmarking} learning algorithms
\end{enumerate}
\end{frame}
\begin{frame}{Trends and Challenges}
Dissimilar Datasets
\begin{enumerate}
\item Quality of Kullback divergence measured on the datasets with different source/target problems
\item TL need to be more robust for \textit{heterogenous} problems
\item Problem gets severe when data features are not \textit{representative}
\item \textit{Big Data} calls for rapid development of TL algorithms
\end{enumerate}
\end{frame}
\bibliographystyle{unsrt}
\nocite{*}
\bibliography{main}
\end{document}
答案1
为了在课堂上显示参考书目，beamer您需要使用以下代码（参见添加的frame！）：
\begin{frame}{Bibliography}
\bibliographystyle{unsrt}
\nocite{*}
\bibliography{\jobname}
\end{frame}
我缩短了您给出的代码，并将缺失的代码添加frame到以下 MWE
\RequirePackage{filecontents}
\begin{filecontents*}{\jobname.bib}
@Book{goossens,
author = {Goossens, Michel and Mittelbach, Frank and
Samarin, Alexander},
title = {The LaTeX Companion},
edition = {1},
publisher = {Addison-Wesley},
location = {Reading, Mass.},
year = {1994},
}
@Book{adams,
title = {The Restaurant at the End of the Universe},
author = {Douglas Adams},
series = {The Hitchhiker's Guide to the Galaxy},
publisher = {Pan Macmillan},
year = {1980},
}
\end{filecontents*}
\documentclass[11pt]{beamer}
\usetheme{Berkeley}
\usepackage[utf8]{inputenc}
\usepackage[english]{babel}
%\usepackage{graphicx} % not needed, loaded by class
\author{Upendra Pratap Singh}
\title{Transfer Learning}
%\setbeamertemplate{navigation symbols}{\insertlogo}
%\setbeamercovered{transparent}
%\setbeamertemplate{navigation symbols}{}
\logo{\includegraphics[height=1.6cm]{example-image-a}} % logo.png
\institute{Indian Institute of Information Technology, Allahabad}
\date{\today}
\subject{Data Analytics and Machine Learning}
\begin{document}
\begin{frame}
\titlepage
\end{frame}
\begin{frame}{Bibliography}
\bibliographystyle{unsrt}
\nocite{*}
\bibliography{\jobname}
\end{frame}
\end{document}
生成的参考书目页面如下：
