**When scientific research is conducted, it gets published for peer review by other scientists in order to validate the findings against errors in the procedures or conclusions derived. *Can an AI take this task over?* Maybe not completely, but machine learning and artificial intelligence are currently automating parts of the review process to help scientists.**

<center><img src="http://i.imgur.com/kGTK1HW.jpg" /><br/><em><a href="https://techxplore.com/news/2017-03-scientists-ai-peer.html">Source: The cover of the March issue of Communications of the ACM</a></em></center>

The top magazine for the Association for Computing Machinery, the *Communications of the ACM*, is [publishing a paper](http://dl.acm.org/citation.cfm?doid=3055102.2979672) about the benefits of using AI to improve the review process even more.

A program called [SubSift](http://subsift.ilrt.bris.ac.uk/) has already been laying inroads to demonstrate the effectiveness of AI to help reviews. Simon Price and Peter Flach developed this system for matching papers with reviewers best suited to do the reviewing of certain papers. It can even pull people to create a peer review panel.

> SubSift, short for *submission sifting*, matches submitted conference or journal papers to potential peer reviewers based on similarity to published works. Conference organisers and journal editors can use the interactive SubSift demonstrations on this site to help support their academic peer review process. - http://subsift.ilrt.bris.ac.uk/

<center><img src="http://i.imgur.com/zA49mva.jpg" /><br/><em><a href="http://subsift.ilrt.bris.ac.uk/">Source: SubSift</a></em></center>

Peter Flach, Professor of Artificial Intelligence in the Department of Computer Science, said: 

*"Our review paper shows that a range of tools for AI-supported peer review already exist and many more are around the corner. We suggest it is now time to rethink the peer review process to make best use of these exciting advances in technology."*

SubSift is open source and hosted on GoogleCode. It was developed in 2009 as a way to assign paper assignments, but was generalized as a web service. It can be used by developers for other software to incorporate this functionality. Another tool called [MLj Matcher](https://github.com/nilesh-c/mljmatcher) was developed in 2010 as an alternative interface for assigning paper journals.

The scientific review process gladly still has humans judging whether to accept or reject scientific publications, but the machine learning and neural networking technology is able to help scientists find papers best suited for their reviewing abilities by revising the plethora of papers out there.

I don't find a problem with this revising process to help get reviews done better, as long as papers aren't being ignored. I just don't want AI to be evaluating papers alone. Humans need to be doing that in the end, even if AI steps in to help more and more.

----
#### Thank you for your time and attention! I appreciate the knowledge reaching more people. Take care. Peace.

---
References:
- [Scientists look to AI for help in peer review](https://techxplore.com/news/2017-03-scientists-ai-peer.html)
- Simon Price et al. Computational support for academic peer review, Communications of the ACM (2017). DOI: [10.1145/2979672](http://dl.acm.org/citation.cfm?doid=3055102.2979672)
- [SubSift](http://subsift.ilrt.bris.ac.uk/)

---
<sub>If you appreciate and value the content, please consider:
Upvoting <img src="http://i.imgur.com/H9LQn1S.png" /> ,&nbsp;&nbsp;&nbsp;&nbsp;Sharing <img src="http://i.imgur.com/ri11ehe.png"  /> **or** &nbsp;&nbsp;Reblogging <img src="http://i.imgur.com/1Abt9EI.png" /> **below**.</sub>

<sub>[![Follow](http://i.imgur.com/aTa4vR8.png)](https://steemit.com/@krnel) me for more content to come!</sub>

<sub>*Looking to contact me?* Find me on [Discord](https://discord.gg/bJCnv5j) or send me a message on [SteemKURE](https://steemkure.com/contact).</sub>

---
<sub>**Please consider supporting me as a [Steem Witness](https://steemit.com/witness-category/@krnel/krnel-s-declaration-of-witness) by voting for me at the [bottom of the Witness page](https://steemit.com/~witnesses); or just click on the upvote button if I am in the top 50:**
<center><img src="http://i.imgur.com/8pdfp0Q.png" /></center></sub>

<sub>If you are unsure how to vote for witnesses, you can put my name in the "**SET PROXY**" section at the bottom of the [Witness Voting page](https://steemit.com/~witnesses) which will use my witness votes.</sub>

---
2017-03-22, 10am