From academic papers
to compelling stories.

Upload your PDFs. We extract key figures, build clear outlines, and craft every slide.

arXiv PDFsConference papersThesis chapters

See It In Action

Watch how Tosea.ai transforms academic papers into compelling presentations in minutes

Experience the power of AI-driven presentation creation

Present your work the way you researched it

Extract figures and formulas from PDFs, shape your story with AI,
and export polished slides—turning reading into presenting.

Evidence-driven Slides

Extract every figure and formula with precision. We show what's actually in your paper—real evidence, not AI hallucinations.

File Content
Deep Residual Learning for Image Recognition
Kaiming He, et al.
Abstract
Deeper neural networks are more difficult to train. We present a residual learning framework to ease the training of networks that are substantially deeper than those used previously. We explicitly reformulate the layers as learning residual functions with reference to the layer inputs, instead of learning unreferenced functions. We provide comprehensive empirical evidence showing that these residual networks are easier to optimize, and can gain accuracy from considerably increased depth. On the ImageNet dataset we evaluate residual nets with a depth of up to 152 layers—8× deeper than VGG nets but still having lower complexity. An ensemble of these residual nets achieves 3.57% error on the ImageNet test set. This result won the 1st place on the ILSVRC 2015 classification task. We also present analysis on CIFAR-10 with 100 and 1000 layers. The depth of representations is of central importance for many visual recognition tasks. Solely due to our extremely deep representations, we obtain a 28% relative improvement on the COCO object detection dataset. Deep residual nets are foundations of our submissions to ILSVRC & COCO 2015 competitions.
CVPR 2016
Images
22/22
Figure 1
Training error
Figure 2
Residual block
Figure 3
Network arch.
Figure 4
Training curves
Figure 5
Comparison
Table 1
Accuracy (%)
Table 2
Error rates
Figure 6
Layer depth

Build Your Outline, Step by Step

Work with your presentation outline—guide AI to add sections, adjust structure, and refine your narrative flow.

Slide 1页
Version 1
Deep Residual Learning
Authors:Kaiming He, et al.
Conference:CVPR 2016
Hello everyone. Today, I'll be presenting the paper 'Deep Residual Learning'...
Slide 2页
Version 1
Presentation Outline
Structure:6 key sections
Here's a quick roadmap for our talk. We'll start by understanding the core problem...
Slide 3页
Version 1
Introduction: Does Deeper Always Mean Better?
Authors:Kaiming He, Xiangyu Zhang
@Edit 1×
Add speaker notes...
Deep Residual Learning for Image Recognition
1. Introduction
Hello everyone. Today, I'll be presenting the paper that not only won the Best Paper Award at CVPR 2016...
2. Presentation Outline
Here's a quick roadmap for our talk. We'll start by understanding the core problem...
3. Introduction: Does Deeper Always Mean Better?
The central question this paper addresses is: is learning better networks as simple as stacking more layers?
4. The Method
The core philosophy of ResNet is brilliantly simple. Instead of forcing layers to learn...
5. The Residual Block Explained
This idea is implemented using what the authors call a 'residual block'. As you can see, the input bypasses the weight layers...
6. Network Architectures: Plain vs. Residual
So, how does this look in a full network? On the left is VGG-19 for reference. In the middle is a 34-layer 'plain' network...
Export to PDF

Perfect Each Slide & Export

Fine-tune pages with AI—adjust layouts only or refine everything—then export presentation-ready files instantly.

Page1页
Layout Only
Download
Before
Deep Residual Learning for Image Recognition
Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun
Microsoft Research
After (Layout Only)
Deep Residual Learning for Image Recognition
Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun
Microsoft Research

Real Results, Real Impact

See what our users have created across disciplines

The replication of DNA in Escherichia coli - Slide 1
1 / 12
BiologyPNAS 1958

The replication of DNA in Escherichia coli

Meselson & Stahl

DNA mechanismExperimental designAuto-extracted
Development of the Colle-Salvetti correlation-energy formula into a functional of the electron density - Slide 1
1 / 12
ChemistryPhys. Rev. B 1988

Development of the Colle-Salvetti correlation-energy formula into a functional of the electron density

Lee, Yang & Parr

Complex equationsTheoretical modelsChemical formulas
Attention is All you Need - Slide 1
1 / 12
Computer ScienceNeurIPS 2017

Attention is All you Need

Vaswani, et al.

Architecture diagramPerformance tablesCode-ready
A Mathematical Theory of Communication - Slide 1
1 / 12
MathematicsBell System 1948

A Mathematical Theory of Communication

Claude Shannon

Mathematical proofsDiagramsFoundational theory
The Colonial Origins of Comparative Development - Slide 1
1 / 12
EconomicsAER 2001

The Colonial Origins of Comparative Development

Acemoglu, Johnson & Robinson

Statistical analysisData visualizationEmpirical results
Streptomycin Treatment of Pulmonary Tuberculosis - Slide 1
1 / 12
MedicineBMJ 1948

Streptomycin Treatment of Pulmonary Tuberculosis

Medical Research Council

Clinical dataTreatment outcomesEvidence-based

Built for Every Research Scenario

From academic conferences to any professional setting

Academic Conference

Present at CVPR, NeurIPS, ICML, and more

Papers have many figures, manual organization is time-consuming
Auto-extract all figures and formulas, generate structured outline

Lab Meeting & Seminar

Weekly literature sharing and group discussions

Preparing paper presentations every week takes hours
Quickly convert papers to slides, support iterative versions

Online Course & Webinar

Record lessons and host webinars

Breaking down textbook content into slides
Auto-split by chapters, add speaker notes

Literature Review

Organize reading notes and research summaries

Reading many papers, hard to organize insights
Extract key information, structured presentation

Client Presentation

Sales demos and consulting presentations

Client materials need consistent formatting
Generate professional PPT, support custom templates

Simple, Transparent Pricing

Choose the plan that fits your needs perfectly

Free

Try it out, get started quickly

USD0

Forever free

Get Started
200 credits/month
6 curated templates
Basic AI models
200 bonus credits on signup
Most Popular

Pro

Professional creation, beautiful presentations

USD149.99

/ year (approx. $12.50/month)

Upgrade to Pro
1,500 credits/month
All templates + premium models
No watermark exports
High-precision extraction

Max

Heavy usage, best value

USD49.99

/ month billed monthly

Upgrade to Max
6,000 credits/month (incl. bonus)
Unlimited exports
Up to 50 slides
Beta features early access

Your Research Deserves a Better Stage

Stop wrestling with slides. Start telling your story.

No installation · Online generation
Support PDF / Markdown / URL
Export PPTX / PDF