This paper presents a object detection algorithm that improves mAP on PASCAL VOC dataset by over 20% to previous state-of-the-art. Unlike image classification which take an image or the center part of an image as input, object detection task requires an algorithm to detect bounding boxes of objects in an image. To use the high capacity CNN features in object detection, the proposed algorithm first generates region proposals. CNN features are extracted from those region proposals and are feed to a set of class-specific linear SVMs which tell whether objects are detected in those regions.
Technical details

The figure below show the object detection system in this paper.
![](http://3.bp.blogspot.com/-O6e43qcpcYA/VWapFWyXt5I/AAAAAAAAA8c/rcjlQJAQ35s/s320/system.png)

Because the PASCAL VOC dataset is not large enough for training high capacity CNN features, this paper use supervised pre-training on a large auxiliary dataset (ILSVRC 2012). The CNN is then fine-tuned with a portion of the PASCAL VOC dataset.

Results

The following table shows the detection mAP on VOC 2007 test.

![](http://1.bp.blogspot.com/-AmEd1cI6iWs/VWaqnD1YYwI/AAAAAAAAA8o/07pfZpdvwck/s400/table%2B2.png)

Are all language orders as hard?
Supposedly, for RNNs yes, for Transformers no

@JenniferCWhite @ryandcotterell
aclanthology.org/2021.acl-long.…

github.com/rycolab/artifi… (currently empty)
#NLProc
Really cool, with a caveat
Adapted from: https://twitter.com/LChoshen/status/1450809889106931716?s=20

GitHub - rycolab/artificial-languages
Contribute to rycolab/artificial-languages development by creating an account on GitHub.
https://github.com/rycolab/artificial-languages
The paper creates synthetic languages (using a PCFG) with various ordering rules, being able to compare each order.
Image
They also add agreement, and a vocabulary to introduce more of real language important features (e.g. long-distance dependencies) 
Last, the words are taken from words that could have been in English (e.g. daxing) 
Their results are the Transformers do have inductive biases, but those are not towards the most probable language orders (e.g. VOS is quite easy for Transformers)
Image
Another interesting thing is that RNNs do not show that. Despite all that we know about their strong inductive bias regarding forgetting (hard to relate far away words\information). 
Last, it seems that a lot of what makes things easy and hard is related to the agreement, at least from looking at which rules matter together.
Image
I really like this direction (which they seem to discuss a lot, seems like a reviewer was quite annoying) and the results are interesting.

However, honestly, I also have a lot of criticism about this work. 
A lot of the details are missing. For example, how were the words tokenized? If BPE, then morphology is split to tokens perfectly? If none, then their sophisticated way of choosing words doesn't matter, because they anyway convert the words into one hot vectors? 
How large were the vocabularies? What were the parameters of the networks train? There are works saying that depth changes inductive biases and not only arch... 
Because of the large number of networks (~128 * 10 reps), they train on really small amounts of data. Hard to say to which amount did that matter, and whether with a small vocab it is reasonable. Still, hard to generalize from the results because of it. 
Last, and that is especially peculiar:
"Weights given to each production were chosen manually through experimentation. Some principles for choosing weights for a grammar in this manner are described by Eisner and Smith (2008)" 
So, we can't know what type of sentences they create (balanced trees, long sentences, really low weights on something that makes a phenomenon\switch less interesting). And they don't share the method or even the chosen weights. 
To sum, this is a very interesting approach, with a lot of potential, that could benefit from more analysis of the results (e.g. per switch), and analysing myself is also problematic as a lot is left out of the paper (nor there is an appendix with the info.)

Although women in both Iraq and Afghanistan have taken part in ground combat and although restrictions have been lifted officially, there remains a strong resistance to the policy that now allows women to work in officially designated combat-related jobs. This resistance is especially strong among online comments to military media coverage. This study examines narratives of masculinity and resistance to women’s presence in combat through online commentary by mostly men who self-identify as military past and present. Using narrative analysis as technique we examine the hegemonic conditions of military culture that render women’s participation problematic. We identify a melodrama, an emotional, simplified, persuasive, and morally laden narrative surrounding the battle between ‘good’ and ‘evil’,one that results in what we term masculinity under attack, complete with prototypical plots, characters, and moral evaluations, such as blameless victims, evil villains, and (blocked) heroes. Because narratives are strategic, persuasive, and deeply ideological, this work offers a unique examination that connects the local narratives of this online community to a broader cultural narrative, one that has material implications for women and all things feminine in the military.

People turn into a flat, hot gas in mosh pits.

The paper introduces a model which is probabilistic for non linear manifold discovery. It is based on a generative model with missing variables and required a variational EM implementation which is standard but nevertheless technical to derive in this specific context.