paper_id
stringlengths 10
10
⌀ | yes_no
bool 2
classes | paper_index
int64 0
519
⌀ | evidence
stringlengths 0
37.7k
| question
stringlengths 4
11.7k
| answer
stringlengths 1
26k
|
|---|---|---|---|---|---|
null | false
| 2
|
There has been significant progress on Named Entity Recognition (NER) in recent years using models based on machine learning algorithms BIBREF0 , BIBREF1 , BIBREF2 . As with other Natural Language Processing (NLP) tasks, building NER systems typically requires a massive amount of labeled training data which are annotated by experts. In real applications, we often need to consider new types of entities in new domains where we do not have existing annotated data. For such new types of entities, however, it is very hard to find experts to annotate the data within short time limits and hiring experts is costly and non-scalable, both in terms of time and money.
In order to quickly obtain new training data, we can use crowdsourcing as one alternative way at lower cost in a short time. But as an exchange, crowd annotations from non-experts may be of lower quality than those from experts. It is one biggest challenge to build a powerful NER system on such a low quality annotated data. Although we can obtain high quality annotations for each input sentence by majority voting, it can be a waste of human labors to achieve such a goal, especially for some ambiguous sentences which may require a number of annotations to reach an agreement. Thus majority work directly build models on crowd annotations, trying to model the differences among annotators, for example, some of the annotators may be more trustful BIBREF3 , BIBREF4 .
Here we focus mainly on the Chinese NER, which is more difficult than NER for other languages such as English for the lack of morphological variations such as capitalization and in particular the uncertainty in word segmentation. The Chinese NE taggers trained on news domain often perform poor in other domains. Although we can alleviate the problem by using character-level tagging to resolve the problem of poor word segmentation performances BIBREF5 , still there exists a large gap when the target domain changes, especially for the texts of social media. Thus, in order to get a good tagger for new domains and also for the conditions of new entity types, we require large amounts of labeled data. Therefore, crowdsourcing is a reasonable solution for these situations.
In this paper, we propose an approach to training a Chinese NER system on the crowd-annotated data. Our goal is to extract additional annotator independent features by adversarial training, alleviating the annotation noises of non-experts. The idea of adversarial training in neural networks has been used successfully in several NLP tasks, such as cross-lingual POS tagging BIBREF6 and cross-domain POS tagging BIBREF7 . They use it to reduce the negative influences of the input divergences among different domains or languages, while we use adversarial training to reduce the negative influences brought by different crowd annotators. To our best knowledge, we are the first to apply adversarial training for crowd annotation learning.
In the learning framework, we perform adversarial training between the basic NER and an additional worker discriminator. We have a common Bi-LSTM for representing annotator-generic information and a private Bi-LSTM for representing annotator-specific information. We build another label Bi-LSTM by the crowd-annotated NE label sequence which reflects the mind of the crowd annotators who learn entity definitions by reading the annotation guidebook. The common and private Bi-LSTMs are used for NER, while the common and label Bi-LSTMs are used as inputs for the worker discriminator. The parameters of the common Bi-LSTM are learned by adversarial training, maximizing the worker discriminator loss and meanwhile minimizing the NER loss. Thus the resulting features of the common Bi-LSTM are worker invariant and NER sensitive.
For evaluation, we create two Chinese NER datasets in two domains: dialog and e-commerce. We require the crowd annotators to label the types of entities, including person, song, brand, product, and so on. Identifying these entities is useful for chatbot and e-commerce platforms BIBREF8 . Then we conduct experiments on the newly created datasets to verify the effectiveness of the proposed adversarial neural network model. The results show that our system outperforms very strong baseline systems. In summary, we make the following contributions:
In order to quickly obtain new training data, we can use crowdsourcing as one alternative way at lower cost in a short time.
|
What is the alternative way to quickly obtain new training data?
|
The crowdsourcing could be used as one alternative way at lower cost in a short time.
|
null | false
| null |
Which of the following are renewable energies and which are fossil fuels: natural gas, wind energy, hydropower, oil shales, solar power, coal, and petroleum.
|
Hydropower, solar power, wind energy are considered to be renewable energies. Coal, petroleum, natural gas, and oil shales are defined as fossil fuels.
|
|
null | false
| null |
Compose a rhyming toast to getting together with old friends
|
Back together once again we are,
Though we now live near and far.
Time has passed and we are still friends,
This is not where our story ends.
|
|
null | false
| null |
How does the sun sustain life on earth?
|
The sun is what sustains life on earth. It provides the light needed for plants to grow through photosynthesis. It is a main source of Vitamin D for humans, and it is a source of energy we can capture on earth and convert into usable electricity through the use of solar panels. It is what causes weather and climate changes on the earth. Animals and plants are completely dependent on the sun because it creates the food we need to survive.
|
|
null | false
| null |
Categorize each of the following as either 'spicy', 'not spicy' or 'sometimes spicy': jalapenos, chilis, sriracha, chips, pizza, cocktails, fruit, milk
|
Jalapenos are spicy.
Chilis are spicy.
Sriracha is spicy.
Chips are sometimes spicy.
Pizza is sometimes spicy.
Cocktails are sometimes spicy.
Fruit is not spicy.
MIlk is not spicy.
|
|
null | false
| null |
Healthy choice for kids or unhealthy: French Fries, Banana, Candy, Vegetables
|
Fruits like Banana and Vegetables are healthy choices for kids. They love fried food and sweets. These should be consumed moderately. French Fries and Candy are unhealthy choices.
|
|
null | false
| 199
|
Evidence-based medicine (EBM) is of primary importance in the medical field. Its goal is to present statistical analyses of issues of clinical focus based on retrieving and analyzing numerous papers in the medical literature BIBREF0 . The PubMed database is one of the most commonly used databases in EBM BIBREF1 .
Biomedical papers, describing randomized controlled trials in medical intervention, are published at a high rate every year. The volume of these publications makes it very challenging for physicians to find the best medical intervention for a given patient group and condition BIBREF2 . Computational methods and natural language processing (NLP) could be adopted in order to expedite the process of biomedical evidence synthesis. Specifically, NLP tasks applied to well structured documents and queries can help physicians extract appropriate information to identify the best available evidence in the context of medical treatment.
Clinical questions are formed using the PIO framework, where clinical issues are broken down into four components: Population/Problem (P), Intervention (I), Comparator (C), and Outcome (O). We will refer to these categories as PIO elements, by using the common practice of merging the C and I categories. In BIBREF3 a literature screening performed in 10 systematic reviews was studied. It was found that using the PIO framework can significantly improve literature screening efficacy. Therefore, efficient extraction of PIO elements is a key feature of many EBM applications and could be thought of as a multi-label sentence classification problem.
Previous works on PIO element extraction focused on classical NLP methods, such as Naive Bayes (NB), Support Vector Machines (SVM) and Conditional Random Fields (CRF) BIBREF4 , BIBREF5 . These models are shallow and limited in terms of modeling capacity. Furthermore, most of these classifiers are trained to extract PIO elements one by one which is sub-optimal since this approach does not allow the use of shared structure among the individual classifiers.
Deep neural network models have increased in popularity in the field of NLP. They have pushed the state of the art of text representation and information retrieval. More specifically, these techniques enhanced NLP algorithms through the use of contextualized text embeddings at word, sentence, and paragraph levels BIBREF6 , BIBREF7 , BIBREF8 , BIBREF9 , BIBREF10 , BIBREF11 .
More recently, jin2018pico proposed a bidirectional long short term memory (LSTM) model to simultaneously extract PIO components from PubMed abstracts. To our knowledge, that study was the first in which a deep learning framework was used to extract PIO elements from PubMed abstracts.
In the present paper, we build a dataset of PIO elements by improving the methodology found in BIBREF12 . Furthermore, we built a multi-label PIO classifier, along with a boosting framework, based on the state of the art text embedding, BERT. This embedding model has been proven to offer a better contextualization compared to a bidirectional LSTM model BIBREF9 .
In the present paper, we build a dataset of PIO elements by improving the methodology found in (Jin and Szolovits, 2018).
|
What dataset did the authors build?
|
A dataset of PIO elements.
|
null | false
| null |
How to escape from a helicopter trapped in water ?
|
If you are ever trapped inside a helicopter while submerged in water, it’s best to try and remain calm until the cabin is completely underwater. It’s better to wait for pressure to be equalized, before you try to open the door or break the glass to escape.
|
|
null | false
| null |
I want to become a pro MMA fighter. How can I go about doing so?
|
There are many different avenues that you can take in order to become a pro Mixed-Martial Arts fighter. To be successful will require a lot of discipline and hard work. Here is one potential path which you may consider:
1) Improve you general fitness: In order to become a successful MMA athlete you need to have good general fitness. This can include things like strength training, yoga (flexibility) and cardio (swimming, running, etc).
2) Get a breadth of experience: It is helpful to get competency in the various specialties that are found in MMA fights so you can adequately handle all styles that may be employed at the time. At a high level, this can be split into two sections:
- Striking: Relevant marital arts you may want to consider training is Boxing, Kickboxing, Muai-Thai.
- Grappling: Relevant marital arts you may want to consider training in is Brazilian Jiu-Jitsu, Sambo, Freestyle Wresting.
3) Consider specializing: While it is good to have a breadth of different techniques, really honing in on one sub-discipline can give you an edge in comparison to other fighters. For example, Khabib Nurmagomedov has found success with a wrestling base.
These are just some starting points to consider when thinking about becoming a professional MMA athlete. Soft Skills like networking and marketing are also very important once you reach more professional levels of the sport.
|
|
null | false
| null |
Electronics deals with electrical circuits that involve active electrical components such as vacuum tubes, transistors, diodes, sensors and integrated circuits, and associated passive interconnection technologies. The nonlinear behaviour of active components and their ability to control electron flows makes digital switching possible, and electronics is widely used in information processing, telecommunications, and signal processing. Interconnection technologies such as circuit boards, electronics packaging technology, and other varied forms of communication infrastructure complete circuit functionality and transform the mixed components into a regular working system.
Today, most electronic devices use semiconductor components to perform electron control. The underlying principles that explain how semiconductors work are studied in solid state physics, whereas the design and construction of electronic circuits to solve practical problems are part of electronics engineering.
|
What is electronics engineering?
|
Electronics engineering is the design and construction of electronic circuits to solve practical problems.
|
null | false
| null |
Choose the word which is different from the rest: Car, Scooter, Helicopter, Aeroplane, Cycle
|
All except Cycle run on fuel.
|
|
null | false
| null |
Classify the following as "animate" or "inanimate": humans, rocks, houses, plants, trees, cars, flowers, curtains, walls
|
Humans: Animate
Rocks: Inanimate
Houses: Inanimate
Plants: Animate
Trees: Animate
Cars: Inanimate
Flowers: Animate
Curtains: Inanimate
Walls: Inanimate
|
|
null | false
| null |
The Class 15E 4-8-2 Mountain type mixed traffic steam locomotive was a refinement of the Classes 15C and 15CA. It was designed by A.G. Watson, Chief Mechanical Engineer (CME) of the South African Railways (SAR) from 1929 to 1936, and incorporated many of the improvements which had been developed by him, some of which were a vastly enlarged standardised boiler, a large and wide fire grate and a Watson cab.
|
Who designed the South African Class 15E 4-8-2 locomotive?
|
The designer of the South African Class 15E 4-8-2 locomotive was A.G. Watson, Chief Mechanical Engineer of the South African Railways.
|
null | false
| 45
|
End-to-end speech translation aims to translate a piece of audio into a target-language translation in one step. The raw speech signals are usually converted to sequences of acoustic features, e.g. Mel filterbank features. Here, we define the speech feature sequence as $\mathbf {x} = (x_1, \cdots , x_{T_x})$.The transcription and translation sequences are denoted as $\mathbf {y^{s}} = (y_1^{s}, \cdots , y_{T_s}^{s})$, and $\mathbf {y^{t}} = (y_1^{t}, \cdots , y_{T_t}^{t})$ repectively. Each symbol in $\mathbf {y^{s}}$ or $\mathbf {y^{t}}$ is an integer index of the symbol in a vocabulary $V_{src}$ or $V_{trg}$ respectively (e.g. $y^s_i=k, k\in [0, |V_{src}|-1]$). In this work, we suppose that an ASR dataset, an MT dataset, and a ST dataset are available, denoted as $\mathcal {A} = \lbrace (\mathbf {x_i}, \mathbf {y^{s}_i})\rbrace _{i=0}^I$, $\mathcal {M} =\lbrace (\mathbf {y^{s}_j}, \mathbf {y^{t}_j})\rbrace _{j=0}^J$ and $ \mathcal {S} =\lbrace (\mathbf {x_l}, \mathbf {y^{t}_l})\rbrace _{l=0}^L$ respectively. Given a new piece of audio $\mathbf {x}$, our goal is to learn an end to end model to generate a translation sentence $\mathbf {y^{t}}$ without generating an intermediate result $\mathbf {y^{s}}$.
To leverage large scale ASR and MT data, multi-task learning and pre-training techniques are widely employed to improve the ST system. As shown in Figure FIGREF4, there are three popular multi-task strategies for ST, including 1) one-to-many setting, in which a speech encoder is shared between ASR and ST tasks; 2) many-to-one setting in which a decoder is shared between MT and ST tasks; and 3) many-to-many setting where both the encoder and decoder are shared.
A many-to-many multi-task model contains two encoders as well as two decoders. It can be jointly trained on ASR, MT, and ST tasks. As the attention module is task-specific, three attentions are defined.
Usually, the size of $\mathcal {A}$ and $\mathcal {M}$ is much larger than $\mathcal {S}$. Therefore, the common training practice is to pre-train the model on ASR and MT tasks and then fine-tune it with a multi-task learning manner. However, as aforementioned, this method suffers from subnet waste, role mismatch and non-pre-trained attention issues, which severely limits the end-to-end ST performance.
Therefore, the common training practice is to pre-train the model on ASR and MT tasks and then fine-tune it with a multi-task learning manner. However, as aforementioned, this method suffers from subnet waste, role mismatch and non-pre-trained attention issues, which severely limits the end-to-end ST performance.
|
Are there any limitations on the common training practice to pre-train the model on ASR and MT tasks and then fine-tune it with a multi-task learning manner?
|
Yes, there are the problems of subnet waste, role mismatch, and non-pre-trained attention issues.
|
null | false
| null |
What are some tricks to learning a new language?
|
Learning a new language can be a great challenge. Here are some tips to help you learn a new language:
1. Turn on Netflix subtitles or dubbing. You'll hear and read how people actually talk, including slang.
2. Join an online chat group for other people learning your language or find someone in your community to practice speaking with.
3. Use flash cards to common words and phrases.
4. Listen to radio or podcasts in your new language.
|
|
null | false
| null |
What is the black oil on the TV show The X-Files?
|
The black oil, also known as purity or the black cancer, is a sentient, fluid alien virus that resembles a black oil slick and is capable of independent movement. It infects humanoids by entering orifices, like eyes and mouths, and gains control of their victim. The alien colonists use the black oil as their reproductive life force.
|
|
null | false
| null |
What is a Limited company?
|
This type of company limits how much its members will have to pay if the company is wound up. The members of most limited companies will only have to pay any money unpaid on their shares. If a company limited by guarantee is wound up, the money its members have to pay is limited to the amount shown in the memorandum.
|
|
null | false
| null |
Identify which instrument is string or woodwind: Giga, Panflute, Chuniri, Dizi
|
Dizi and Panflute are woodwind, Giga and Chuniri are string.
|
|
1703.04617
| false
| null |
Enabling computers to understand given documents and answer questions about their content has recently attracted intensive interest, including but not limited to the efforts as in BIBREF0 , BIBREF1 , BIBREF2 , BIBREF3 , BIBREF4 , BIBREF5 . Many specific problems such as machine comprehension and question answering often involve modeling such question-document pairs.
machine comprehension
Nelufar
|
What MC abbreviate for?
|
machine comprehension
|
null | false
| 397
|
The goal of meeting transcription is to have machines generate speaker-annotated transcripts of natural meetings based on their audio and optionally video recordings. Meeting transcription and analytics would be a key to enhancing productivity as well as improving accessibility in the workplace. It can also be used for conversation transcription in other domains such as healthcare BIBREF0. Research in this space was promoted in the 2000s by NIST Rich Transcription Evaluation series and public release of relevant corpora BIBREF1, BIBREF2, BIBREF3. While systems developed in the early days yielded high error rates, advances have been made in individual component technology fields, including conversational speech recognition BIBREF4, BIBREF5, far-field speech processing BIBREF6, BIBREF7, BIBREF8, and speaker identification and diarization BIBREF9, BIBREF10, BIBREF11. When cameras are used in addition to microphones to capture the meeting conversations, speaker identification quality could be further improved thanks to the computer vision technology. These trends motivated us to build an end-to-end audio-visual meeting transcription system to identify and address unsolved challenges. This report describes our learning, with focuses on overall architecture design, overlapped speech recognition, and audio-visual speaker diarization.
When designing meeting transcription systems, different constraints must be taken into account depending on targeted scenarios. In some cases, microphone arrays are used as an input device. If the names of expected meeting attendees are known beforehand, the transcription system should be able to provide each utterance with the true identity (e.g., “Alice” or “Bob”) instead of a randomly generated label like “Speaker1”. It is often required to show the transcription in near real time, which makes the task more challenging.
This work assumes the following scenario. We consider a scheduled meeting setting, where an organizer arranges a meeting in advance and sends invitations to attendees. The transcription system has access to the invitees' names. However, actual attendees may not completely match those invited to the meeting. The users are supposed to enroll themselves in the system beforehand so that their utterances in the meeting can be associated with their names. The meeting is recorded with an audio-visual device equipped with a seven-element circular microphone array and a fisheye camera. Transcriptions must be shown with a latency of up to a few seconds.
This paper investigates three key challenges.
Speech overlaps: Recognizing overlapped speech has been one of the main challenges in meeting transcription with limited tangible progress. Numerous multi-channel speech separation methods were proposed based on independent component analysis or spatial clustering BIBREF12, BIBREF13, BIBREF14, BIBREF15, BIBREF16, BIBREF17. However, there was little successful effort to apply these methods to natural meetings. Neural network-based single-channel separation methods using techniques like permutation invariant training (PIT) BIBREF18 or deep clustering (DC) BIBREF19 are known to be vulnerable to various types of acoustic distortion, including reverberation and background noise BIBREF20. In addition, these methods were tested almost exclusively on small-scale segmented synthetic data and have not been applied to continuous conversational speech audio. Although the recently held CHiME-5 challenge helped the community make a step forward to a realistic setting, it still allowed the use of ground-truth speaker segments BIBREF21, BIBREF22.
We address this long-standing problem with a continuous speech separation (CSS) approach, which we proposed in our latest conference papers BIBREF23, BIBREF24. It is based on an observation that the maximum number of simultaneously active speakers is usually limited even in a large meeting. According to BIBREF25, two or fewer speakers are active for more than 98% of the meeting time. Thus, given continuous multi-channel audio observation, we generate a fixed number, say $N$, of time-synchronous signals. Each utterance is separated from overlapping voices and background noise. Then, the separated utterance is spawned from one of the $N$ output channels. For periods where the number of active speakers is fewer than $N$, the extra channels generate zeros. We show how continuous speech separation can fit in with an overall meeting transcription architecture to generate speaker-annotated transcripts.
Note that our speech separation system does not make use of a camera signal. While much progress has been made in audio-visual speech separation, the challenge of dealing with all kinds of image variations remains unsolved BIBREF26, BIBREF27, BIBREF28.
Extensible framework: It is desirable that a single transcription system be able to support various application settings for both maintenance and scalability purposes. While this report focuses on the audio-visual setting, our broader work covers an audio-only setting as well as the scenario where no prior knowledge of meeting attendees is available. A modular and versatile architecture is desired to encompass these different settings.
To this end, we propose a framework called SRD, which stands for “separate, recognize, and diarize”, where CSS, speech recognition, and speaker diarization takes place in tandem. Performing CSS at the beginning allows the other modules to operate on overlap-free signals. Diarization is carried out after speech recognition because its implementation can vary significantly depending on the application settings. By choosing an appropriate diarization module for each setting, multiple use cases can be supported without changing the rest of the system. This architecture also allows transcriptions to be displayed in real time without speaker information. Speaker identities for each utterance may be shown after a couple of seconds.
Audio-visual speaker diarization: Speaker diarization, a process of segmenting input audio and assigning speaker labels to the individual segments, can benefit from a camera signal. The phenomenal improvements that have been made to face detection and identification algorithms by convolutional neural networks (CNNs) BIBREF29, BIBREF30, BIBREF31 make the camera signal very appealing for speaker diarization. While much prior work assumes the batch processing scenario where the entire meeting recording can be processed multiple times, several studies deal with online processing BIBREF32, BIBREF33, BIBREF34, BIBREF35. However, no previous studies comprehensively address the challenges that one might encounter in real meetings. BIBREF32, BIBREF33 do not cope with speech overlaps. While the methods proposed in BIBREF34, BIBREF35 address the overlap issue, they rely solely on spatial cues and thus are not applicable when multiple speakers sit side by side.
Our diarization method handles overlapping utterances as well as co-located speakers by utilizing the time-frequency (TF) masks generated by CSS in speaker identification and sound source localization (SSL). In addition, several enhancements are made to face identification to improve robustness to image variations caused by face occlusions, extreme head pose, lighting conditions, and so on.
We consider a scheduled meeting setting, where an organizer arranges a meeting in advance and sends invitations to attendees.
|
What is the scheduled meeting setting in the system?
|
It means an organizer arranges a meeting in advance and sends invitations to attendees.
|
null | false
| null |
What is the best way to build relationships?
|
The best way to build relationships is to show people your most true authentic self. Communicate clearly and honestly with integrity. Show that you care for those people that you have strong relationships with by listening to them, giving honest opinions, and sharing your experiences openly with them.
|
|
null | false
| null |
Give me a list of things I can do with my apple iwatch
|
There are a lot of things you can potentially do with an apple iwatch. For starters, if you're into fitness, you can use it to monitor your fitness and weight training goals. Additionally, you can also use it to monitor your sleep cycle to ensure that you are getting the best sleep possible. To learn more about the iwatch, go to the apple website and read about it full list of features.
|
|
null | false
| null |
Would each of the following conditions be good or bad for a bike ride? Pouring rain, sunny skies, hail, cloudy day, dense fog, heavy wind, light breeze.
|
Pouring rain: bad, sunny skies: good, hail: bad, cloudy day: good, dense fog: bad, heavy wind: bad, light breeze: good.
|
|
null | false
| null |
In photography, exposure value (EV) is a number that represents a combination of a camera's shutter speed and f-number, such that all combinations that yield the same exposure have the same EV (for any fixed scene luminance). Exposure value is also used to indicate an interval on the photographic exposure scale, with a difference of 1 EV corresponding to a standard power-of-2 exposure step, commonly referred to as a stop.
The EV concept was developed by the German shutter manufacturer Friedrich Deckel in the 1950s (Gebele 1958; Ray 2000, 318). Its intent was to simplify choosing among equivalent camera exposure settings by replacing combinations of shutter speed and f-number (e.g., 1/125 s at f/16) with a single number (e.g., 15). On some lenses with leaf shutters, the process was further simplified by allowing the shutter and aperture controls to be linked such that, when one was changed, the other was automatically adjusted to maintain the same exposure. This was especially helpful to beginners with limited understanding of the effects of shutter speed and aperture and the relationship between them. But it was also useful for experienced photographers who might choose a shutter speed to stop motion or an f-number for depth of field, because it allowed for faster adjustment—without the need for mental calculations—and reduced the chance of error when making the adjustment.
|
What are the main functions on a DSLR camera that affect the correct the exposure of an image?
|
If a DSLR camera is not set to 'Automatic' mode, there are 3 primary functions that need to be considered when deciding on the correct exposure for your image.
1) the Aperture - how wide the aperture on the lens is set determines how much light can pass through the aperture. A small aperture (e.g. f/16) allows less light to reach the sensor as opposed to a wide aperture (e.g f/2.4).
2) the Shutter Speed - how long the shutter is allowed to remain open after you depress the release on the camera. A short exposure time (e.g. 1/1000 sec) will allow less light to hit the cameras sensor whilst a long exposure (e.g. 1/4 sec) will let more light hit the sensor
3) the ISO setting - this setting is used to control how sensitive the sensor is to light hitting it. A low ISO setting (e.g. ISO 200) is much less sensitive than a high ISO setting (e.g. ISO 1600) which means less light is required to capture the for the image on the sensor.
All 3 of these functions work in tandem to ensure the correct exposure value is created based on the image and effect you are trying to create.
|
null | false
| null |
Terror Mountain is a 1928 American silent Western film directed by Louis King and written by Frank Howard Clark and Helen Gregg. The film stars Tom Tyler, Jane Reid, Al Ferguson, Jules Cowles and Frankie Darro. The film was released on August 19, 1928, by Film Booking Offices of America. It was also released as Terror and Tom's Vacation.
|
Who directed Terror Mountain, a American silent Western film?
|
Louis King is a director of the film.
|
null | false
| null |
What was The Last Dance about?
|
The Netflix sports documentary The Last Dance follows the Chicago Bulls as they contest the 1997-1998 NBA season, primarily focusing on the fact this was the last season Michael Jordan played with the team. This documentary made use of documentary footage filmed throughout the season but also tells the life story of Michael Jordan, with a heavy focus on his basketball career. Michael Jordan is widely considered one of the greatest basketball players to have ever lived and as a result the documentary was considered a critical and commercial success for Netflix.
|
|
null | false
| null |
Who was one of the best hip hop groups of the late 90s
|
TLC was one of the most famous girl groups with major success around songs such as waterfalls, no scrubs & unpretty
|
|
1803.07771
| false
| null |
We address the above issues with a new methodology. First, we introduce a two-stage labeling strategy for sentiment texts. In the first stage, annotators are invited to label a large number of short texts with relatively pure sentiment orientations. Each sample is labeled by only one annotator. In the second stage, a relatively small number of text samples with mixed sentiment orientations are annotated, and each sample is labeled by multiple annotators. Second, we propose a two-level long short-term memory (LSTM) BIBREF4 network to achieve two-level feature representation and classify the sentiment orientations of a text sample to utilize two labeled data sets. Lastly, in the proposed two-level LSTM network, lexicon embedding is leveraged to incorporate linguistic features used in lexicon-based methods.
First, we introduce a two-stage labeling strategy for sentiment texts. In the first stage, annotators are invited to label a large number of short texts with relatively pure sentiment orientations. Each sample is labeled by only one annotator. In the second stage, a relatively small number of text samples with mixed sentiment orientations are annotated, and each sample is labeled by multiple annotators.
|
What is the new labeling strategy?
|
They use a two-stage labeling strategy where in the first stage single annotators label a large number of short texts with relatively pure sentiment orientations and in the second stage multiple annotators label few text samples with mixed sentiment orientations
|
null | false
| null |
Identify which instrument is string or percussion: Dhaa, Tremoloa
|
Tremoloa is string, Dhaa is percussion.
|
|
null | false
| null |
why is hydration so important?
|
Water is very good for you because it helps to keep you hydrated. Your body needs water for nearly every function. When you are well hydrated you will look younger because your skin has better elasticity. You can think more clearly because your brain is hydrated. You muscles will recover faster, and you will have better digestion. Coffee and sugar drinks will dehydrate you so if you have some of these make sure you replenish your fluids with more water to get hydrated.
|
|
null | false
| 147
|
We train NMT with RAT to achieve better query translations. We improve a recently proposed NMT baseline, Transformer, that achieves state-of-the-art results for sentence pairs in some languages BIBREF8 . We discuss Transformer, RAT, and our multi-task learning architecture that achieves balanced translation.
We consider two data sources for learning NMT and RAT jointly. The first one is a sentence-level parallel corpus, which we refer to as translation corpus, T C = {(si , ti);i = 1, 2, . . . m}. The second one is the retrieval corpus, which is a collection of k documents RC = {D1, D2, . . . Dk} in the same language as ti .
|
What are the data sources for learning NMT and RAT jointly?
|
The first one is a sentence-level parallel corpus and the second one is the retrieval corpus.
|
null | false
| null |
Is a golden doodle a good pet for a family with young children?
|
Golden doodles, which are part golden retriever and part poodle, are very popular and make good pets, especially for families with children.
While some golden doodles may be outliers, generally they have a very good temperament, and are loving and loyal pets. A young child can yank on their tails or fur, and parents don’t need to be concerned that their child will get bitten (although the golden doodle will likely nudge the child to stop that behavior).
They also are generally intelligent dogs, and learn easily. More importantly, they are easy to house train, which is important for a family with young children, since there aren’t additional messes to clean up after. They don’t shed like long-haired dogs, where the hairs can get on your clothes. They shed in clumps of hair, which are easier to see and clean up.
Last, golden doodles are very cute dogs, which is another element that has contributed to their popularity.
However, because they are part poodle, they can be excitable, in particular when they are puppies, and they need to get a lot of exercise so that they can expend their energy. As a result, they should be with an active family so that they can get plenty of activity.
|
|
null | false
| null |
What are a few things to do when visiting Northern Vietnam?
|
Here are some to do when visiting Northern Vietnam:
1. Go hiking through the picturesque rice terraces in Sa Pa. This is a great way to experience the beauty and local culture of Northern Vietnam.
2. Take a cruise on Ha Long Bay. This Unesco world heritage site is a great place to relax and take in the views as you sail through thousands of limestone islands.
3. Visit Hanoi and go on a food tour. This is one of the best ways to experience the local culture of Vietnam through authentic Vietnamese cuisine.
4. Spend a day exploring the stunning landscapes of Ninh Binh. Home to the ancient capital of Vietnam and the Unesco World Heritage site of Trang An, Ninh Binh is a popular destination in Northern Vietnam.
|
|
null | false
| null |
The Byzantine Empire,[note 1] also referred to as the Eastern Roman Empire or Byzantium, was the continuation of the Roman Empire primarily in its eastern provinces during Late Antiquity and the Middle Ages, when its capital city was Constantinople. It survived the fragmentation and fall of the Western Roman Empire in the 5th century AD and continued to exist for an additional thousand years until the fall of Constantinople to the Ottoman Empire in 1453. During most of its existence, the empire remained the most powerful economic, cultural, and military force in Europe. The terms "Byzantine Empire" and "Eastern Roman Empire" were coined after the end of the realm; its citizens continued to refer to their empire as the Roman Empire and to themselves as Romans[note 2]—a term which Greeks continued to use for themselves into Ottoman times. Although the Roman state continued and its traditions were maintained, modern historians prefer to differentiate the Byzantine Empire from Ancient Rome as it was centered on Constantinople instead of Rome, oriented towards Greek rather than Latin culture, and was characterized by Eastern Orthodox Christianity.
During the high period of the Roman Empire known as the Pax Romana, the western parts of the empire went through Latinization, while the eastern parts of the empire maintained to a large degree their Hellenistic culture. Several events from the 4th to 6th centuries mark the period of transition during which the Roman Empire's Greek East and Latin West diverged. Constantine I (r. 324–337) reorganized the empire, made Constantinople the capital, and legalized Christianity. Under Theodosius I (r. 379–395), Christianity became the state religion, and other religious practices were proscribed. In the reign of Heraclius (r. 610–641), the empire's military and administration were restructured, and Greek was gradually adopted for official use in place of Latin.
The borders of the empire fluctuated through several cycles of decline and recovery. During the reign of Justinian I (r. 527–565), the empire reached its greatest extent after the fall of the west, re-conquering much of the historically Roman western Mediterranean coast, including Africa, Italy and Rome, which it held for two more centuries. The Byzantine–Sasanian War of 602–628 exhausted the empire's resources, and during the early Muslim conquests of the 7th century, it lost its richest provinces, Egypt and Syria, to the Rashidun Caliphate. It then lost Africa to the Umayyads in 698, before the empire was rescued by the Isaurian dynasty.
During the Macedonian dynasty (9th–11th centuries), the empire expanded again and experienced the two-century-long Macedonian Renaissance, which came to an end with the defeat by Seljuk Turks at the Battle of Manzikert in 1071. Civil wars and the ensuing Seljuk invasion led to the loss of most of Asia Minor. The empire recovered during the Komnenian restoration, and by the 12th century, Constantinople was the largest and wealthiest city in Europe.
The empire was delivered a mortal blow during the Fourth Crusade when Constantinople was sacked in 1204 and the territories that the empire formerly governed were divided into competing Byzantine Greek and Latin realms. Despite the eventual recovery of Constantinople in 1261, the Byzantine Empire remained only one of several small rival states in the area for the final two centuries of its existence. Its remaining territories were progressively annexed by the Ottomans in the Byzantine–Ottoman wars over the 14th and 15th centuries.
The fall of Constantinople to the Ottoman Empire in 1453 marked the end of the Byzantine Empire. Refugees fleeing the city after its capture would settle in Italy and other parts of Europe, helping to ignite the Renaissance. The Empire of Trebizond was conquered eight years later when its eponymous capital surrendered to Ottoman forces after it was besieged in 1461. The last Byzantine rump state, the Principality of Theodoro, was conquered by the Ottomans in 1475. Arguments can be made that the fall of the Byzantine Empire to the Ottomans is one of several factors contributing to the end of the Middle Ages and the start of the early modern period.[not verified in body]
|
Given this background text about the Byzantine Empire, which empire can the Byzantines trace their origins?
|
The Byzantine Empire traces it's origins to the Roman Empire
|
null | false
| null |
The Death of Socrates was painted by whom?
|
Jacques Louis David
|
|
null | false
| 160
|
Learning the distributed representation for long spans of text from its constituents has been a key step for various natural language processing (NLP) tasks, such as text classification BIBREF0 , BIBREF1 , semantic matching BIBREF2 , BIBREF3 , and machine translation BIBREF4 . Existing deep learning approaches take a compositional function with different forms to compose word vectors recursively until obtaining a sentential representation. Typically, these compositional functions involve recurrent neural networks BIBREF5 , BIBREF6 , convolutional neural networks BIBREF7 , BIBREF8 , and tree-structured neural networks BIBREF9 , BIBREF10 .
Among these methods, tree-structured neural networks (Tree-NNs) show theirs superior performance in many NLP tasks BIBREF11 , BIBREF12 . Following the syntactic tree structure, Tree-NNs assign a fixed-length vector to each word at the leaves of the tree, and combine word and phrase pairs recursively to create intermediate node vectors, eventually obtaining one final vector to represent the whole sentence.
However, these models have a major limitation in their inability to fully capture the richness of compositionality BIBREF13 . The same parameters are used for all kinds of semantic compositions, even though the compositions have different characteristics in nature. For example, the composition of the adjective and the noun differs significantly from the composition of the verb and the noun. Moreover, many semantic phenomena, such as semantic idiomaticity or transparency, call for more powerful compositional mechanisms BIBREF14 . Therefore, Tree-NNs suffer from the underfitting problem.
To alleviate this problem, some researchers propose to use multiple compositional functions, which are arranged beforehand according to some partition criterion BIBREF11 , BIBREF13 , BIBREF15 . Intuitively, using different parameters for different types of compositions has the potential to greatly reduce underfitting. BIBREF13 [ BIBREF13 ] defined different compositional functions in terms of syntactic categories, and a suitable compositional function is selected based on the syntactic categories. BIBREF15 [ BIBREF15 ] introduced multiple compositional functions and during compositional phase, a proper one is selected based on the input information. Although these models accomplished their mission to a certain extent, they still suffer from the following three challenges. First, the predefined compositional functions cannot cover all the compositional rules; Second, they require more learnable parameters, suffering from the problem of overfitting; Third, it is difficult to determine a universal criterion for semantic composition based solely on syntactic categories.
In this paper, we propose dynamic compositional neural networks over tree structure, in which a meta network is used to generate the context-specific parameters of a dynamic compositional network. Specifically, we construct our models based on two kinds of tree-structured neural networks: recursive neural network (Tree-RecNN) BIBREF11 and tree-structure long short-term memory neural network (Tree-LSTM) BIBREF9 . Our work is inspired by recent work on dynamic parameter prediction BIBREF16 , BIBREF17 , BIBREF18 . The meta network is used to extract the shared meta-knowledge across different compositional rules and to dynamically generate the context-specific compositional function. Thus, the compositional function of our models varies with positions, contexts and samples. The dynamic compositional network then applies those context-specific parameters to the current input information. Both meta and dynamic networks are differentiable such that the overall networks can be trained in an end-to-end fashion. Additional, to reduce the complexity of the whole networks, we define the dynamic weight matrix in a manner simulating low-rank matrix decomposition.
We evaluate our models on two typical tasks: text classification and text semantic matching. The results show that our models are more expressive due to their learning to learn nature, yet without increasing the number of model's parameters. Moreover, we find certain composition operations can be learned implicitly by meta TreeNN, such as the composition of noun phrases and verb phrases.
The contributions of the paper can be summed up as follows.
The results show that our models are more expressive due to their learning to learn nature, yet without increasing the number of model's parameters.
|
Why do the results show that the authors' models are more expressive?
|
Due to their learning to learn nature, yet without increasing the number of model's parameters.
|
null | false
| null |
Give me a list of the best French rugby players ever
|
Antoine Dupont, Damian Penaud, Vincent Clerc, Fabien Pelous, Thierry Dusautoir, Serge Betsen, Fabien Galthié, Romain N'tamack, Cyril Baille, François Cros, Aurélien Rougerie, Nicolas Mas, William Servat
|
|
null | false
| null |
Who has hosted Saturday Night Live the most in the history of the show?
|
Actor Alec Baldwin has hosted Saturday Night Live 17 times since 1990.
|
|
1708.04557
| false
| null |
To estimate speakers' position we use Wordscore BIBREF1 – a version of the Naive Bayes classifier that is deployed for text categorization problems BIBREF22 . In a similar application, BIBREF1 have already demonstrated that Wordscore can be effectively used to derive estimates of TDs policy positions. As in the example above, we pre-process documents by removing all numbers and interjections.
As in the example above, we pre-process documents by removing all numbers and interjections.
|
what processing was done on the speeches before being parsed?
|
Remove numbers and interjections
|
null | false
| null |
How do I make good pour over coffee?
|
Start with a good pour over filter holder, preferably one made of borosilicate glass, and filters especially designed for pour over coffee, such as cone filters. Use a medium-coarse grind of coffee beans, and use filtered water instead of regular tap water. The recommended ratio for coffee to water is 1:15, which means for every 1 gram of coffee, use 15 grams of water. Pour the water over the ground coffee in a circular motion, with a first pour just wetting the coffee grounds, and a second pour slow and steady until to reach the correct ratio. Sip and enjoy.
|
|
null | false
| 17
|
Following developing news stories is imperative to making real-time decisions on important political and public safety matters. Given the abundance of media providers and languages, this endeavor is an extremely difficult task. As such, there is a strong demand for automatic clustering of news streams, so that they can be organized into stories or themes for further processing. Performing this task in an online and efficient manner is a challenging problem, not only for newswire, but also for scientific articles, online reviews, forum posts, blogs, and microblogs.
A key challenge in handling document streams is that the story clusters must be generated on the fly in an online fashion: this requires handling documents one-by-one as they appear in the document stream. In this paper, we provide a treatment to the problem of online document clustering, i.e. the task of clustering a stream of documents into themes. For example, for news articles, we would want to cluster them into related news stories.
To this end, we introduce a system which aggregates news articles into fine-grained story clusters across different languages in a completely online and scalable fashion from a continuous stream. Our clustering approach is part of a larger media monitoring project to solve the problem of monitoring massive text and TV/Radio streams (speech-to-text). In particular, media monitors write intelligence reports about the most relevant events, and being able to search, visualize and explore news clusters assists in gathering more insight about a particular story. Since relevant events may be spawned from any part of the world (and from many multilingual sources), it becomes imperative to cluster news across different languages.
In terms of granularity, the type of story clusters we are interested in are the group of articles which, for example : (i) Narrate recent air-strikes in Eastern Ghouta (Syria); (ii) Describe the recent launch of Space X's Falcon Heavy rocket.
A key challenge in handling document streams is that the story clusters must be generated on the fly in an online fashion: this requires handling documents one-by-one as they appear in the document stream.
|
What is the key challenge in handling document streams?
|
The story clusters must be generated on the fly in an online fashion: this requires handling documents one-by-one as they appear in the document stream.
|
null | false
| null |
How many books are there in the Harry Potter series?
|
7
|
|
null | false
| null |
Is tap water safe to drink in Japan?
|
Yes, tap water is safe to drink in Japan.
|
|
1612.05310
| true
| null |
Reddit is popular website that allows registered users (without identity verification) to participate in fora grouped by topic or interest. Participation consists of posting stories that can be seen by other users, voting stories and comments, and comments in the story's comment section, in the form of a forum. The forums are arranged in the form of a tree, allowing nested conversations, where the replies to a comment are its direct responses. We collected all comments in the stories' conversation in Reddit that were posted in August 2015. Since it is infeasible to manually annotate all of the comments, we process this dataset with the goal of extracting threads that involve suspected trolling attempts and the direct responses to them. To do so, we used Lucene to create an inverted index from the comments and queried it for comments containing the word “troll” with an edit distance of 1 in order to include close variations of this word, hypothesizing that such comments would be reasonable candidates of real trolling attempts. We did observe, however, that sometimes people use the word troll to point out that another user is trolling. Other times, people use the term to express their frustration about a particular user, but there is no trolling attempt. Yet other times people simply discuss trolling and trolls without actually observing one. Nonetheless, we found that this search produced a dataset in which 44.3% of the comments are real trolling attempts. Moreover, it is possible for commenters to believe that they are witnessing a trolling attempt and respond accordingly even where there is none due to misunderstanding. Therefore, the inclusion of comments that do not involve trolling would allow us to learn what triggers a user's interpretation of trolling when it is not present and what kind of response strategies are used.
We had two human annotators who were trained on snippets (i.e., (suspected trolling attempt, responses) pairs) taken from 200 conversations and were allowed to discuss their findings. After this training stage, we asked them to independently label the four aspects for each snippet. We recognize that this limited amount of information is not always sufficient to recover the four aspects we are interested in, so we give the annotators the option to discard instances for which they couldn't determine the labels confidently. The final annotated dataset consists of 1000 conversations composed of 6833 sentences and 88047 tokens. The distribution over the classes per trolling aspect is shown in the table TABREF19 in the column “Size”.
Since it is infeasible to manually annotate all of the comments, we process this dataset with the goal of extracting threads that involve suspected trolling attempts and the direct responses to them.
We had two human annotators who were trained on snippets (i.e., (suspected trolling attempt, responses) pairs) taken from 200 conversations and were allowed to discuss their findings. After this training stage, we asked them to independently label the four aspects for each snippet.
|
Do they use a crowdsourcing platform for annotation?
|
No.
|
1912.07976
| false
| null |
To comprehensive evaluate the performance of the proposed model, the experiments were conducted in three most commonly used ABSA datasets, the Laptops and Restaurant datasets of SemEval-2014 Task4 subtask2 BIBREF0 and an ACL Twitter social dataset BIBREF34. To evaluate our model capability with processing the Chinese language, we also tested the performance of LCF-ATEPC on four Chinese comment datasets BIBREF35, BIBREF36, BIBREF29 (Car, Phone, Notebook, Camera). We preprocessed the seven datasets. We reformatted the origin dataset and annotated each sample with the IOB labels for ATE task and polarity labels for APC tasks, respectively. The polarity of each aspect on the Laptops, Restaurants and datasets may be positive, neutral, and negative, and the conflicting labels of polarity are not considered. The reviews in the four Chinese datasets have been purged, with each aspect may be positive or negative binary polarity. To verify the effectiveness and performance of LCF-ATEPC models on multilingual datasets, we built a multilingual dataset by mixing the 7 datasets. We adopt this dataset to conduct multilingual-oriented ATE and APC experiments.
To evaluate our model capability with processing the Chinese language, we also tested the performance of LCF-ATEPC on four Chinese comment datasets BIBREF35, BIBREF36, BIBREF29 (Car, Phone, Notebook, Camera).
|
In what four Chinese review datasets does LCF-ATEPC achieves state of the art?
|
The answers are shown as follows:
* Car, Phone, Notebook, Camera
|
null | false
| null |
Known as "The Mad" Baldwin, he turned pro in 1970 and lost a decision to Marvin Hagler in 1975. In 1977 he took on Rocky Mosley Jr in the ill-fated U.S. Championship Tournament, but lost via K.O. In 1978 he lost a decision to Marvin Johnson, and retired a year later.
|
Who was boxer John Baldwin
|
Born in Detroit, John Baldwin was a professional US boxer. He fought in the Light-Middleweight category and won a bronze medal in the 1968 Olympics. He turned pro in 1970, and lost a decision to Marvin Hagler in 1975. He retired in 1979
|
null | false
| null |
Should I buy a gravel bike or a down country mountain bike?
|
Although there is reasonable overlap in what you can use gravel and downcountry bikes for, the answer her largely depends on what you plan to do the most of and if trail riding will be important for you. Gravel bikes are great for covering offroad miles more efficiently and long days in the saddle both on road or offroad. However, the lack of any meaningful suspension means that or but a small amount of time off paths and more established trails will be hard work. Downcountry bikes will enable you to go on a wider range of terrain, inlcuding the ability to do trail riding (be very careful before trying to navigate downhill trails on a gravel bike!). However, downcountry bikes are usually heavier but the full suspension and geometry makes them less efficient compared to a gravel bike.
|
|
null | false
| null |
The Sydney Opera House is a multi-venue performing arts centre in Sydney. Located on the foreshore of Sydney Harbour, it is widely regarded as one of the world's most famous and distinctive buildings and a masterpiece of 20th-century architecture.
Designed by Danish architect Jørn Utzon, but completed by an Australian architectural team headed by Peter Hall, the building was formally opened by Queen Elizabeth II on 20 October 1973 after a gestation beginning with Utzon's 1957 selection as winner of an international design competition. The Government of New South Wales, led by the premier, Joseph Cahill, authorised work to begin in 1958 with Utzon directing construction. The government's decision to build Utzon's design is often overshadowed by circumstances that followed, including cost and scheduling overruns as well as the architect's ultimate resignation.
The building and its surrounds occupy the whole of Bennelong Point on Sydney Harbour, between Sydney Cove and Farm Cove, adjacent to the Sydney central business district and the Royal Botanic Gardens, and near to the Sydney Harbour Bridge.
|
When was the Sydney Opera House formally opened
|
The Sydney Opera House is a multi-venue performing arts centre in Sydney. Located on the foreshore of Sydney Harbour, it is widely regarded as one of the world's most famous and distinctive buildings and a masterpiece of 20th-century architecture.
Designed by Danish architect Jorn Utzon, but completed by an Australian architectural team headed by Peter Hall, the building was formally opened by Queen Elizabeth II on 20 October 1973 after a gestation beginning with Utzon's 1957 selection as winner of an international design competition. The Government of New South Wales, led by the premier, Joseph Cahill, authorised work to begin in 1958 with Utzon directing construction. The government's decision to build Utzon's design is often overshadowed by circumstances that followed, including cost and scheduling overruns as well as the architect's ultimate resignation.
The building and its surrounds occupy the whole of Bennelong Point on Sydney Harbour, between Sydney Cove and Farm Cove, adjacent to the Sydney central business district and the Royal Botanic Gardens, and near to the Sydney Harbour Bridge.
|
null | false
| null |
Why is the Saguaro cactus important to Arizona?
|
The Saguaro cactus (Saguaro Cacti) is a very rare type of cacti that only grows in the Sonoran Desert, which southern Arizona is part of. The Saguaro, can easily live a few hundred years as it takes about 50-75 years just to grow one arm. Due to new construction in Arizona, a Saguaro interfering with the location of a building must be transplanted and cannot be destroyed. The destruction of a Saguaro cactus is a class-4 felony.
|
|
null | false
| 172
|
Legal documents are a rather heterogeneous class, which also manifests in their linguistic properties, including the use of named entities and references. Their type and frequency varies significantly, depending on the text type. Texts belonging to specific text type, which are to be selected for inclusion in a corpus must contain enough different named entities and references and they need to be freely available. When comparing legal documents such as laws, court decisions or administrative regulations, decisions are the best option. In laws and administrative regulations, the frequencies of person, location and organization are not high enough for NER experiments. Court decisions, on the other hand, include person, location, organization, references to law, other decision and regulation.
Court decisions from 2017 and 2018 were selected for the dataset, published online by the Federal Ministry of Justice and Consumer Protection. The documents originate from seven federal courts: Federal Labour Court (BAG), Federal Fiscal Court (BFH), Federal Court of Justice (BGH), Federal Patent Court (BPatG), Federal Social Court (BSG), Federal Constitutional Court (BVerfG) and Federal Administrative Court (BVerwG).
From the table of contents, 107 documents from each court were selected (see Table ). The data was collected from the XML documents, i. e., it was extracted from the XML elements Mitwirkung, Titelzeile, Leitsatz, Tenor, Tatbestand, Entscheidungsgründe, Gründen, abweichende Meinung, and sonstiger Titel. The metadata at the beginning of the documents (name of court, date of decision, file number, European Case Law Identifier, document type, laws) and those that belonged to previous legal proceedings was deleted. Paragraph numbers were removed. The extracted data was split into sentences, tokenised using SoMaJo BIBREF16 and manually annotated in WebAnno BIBREF17.
The annotated documents are available in CoNNL-2002. The information originally represented by and through the XML markup was lost in the conversion process. We decided to use CoNNL-2002 because our primary focus was on the NER task and experiments. CoNNL is one of the best practice formats for NER datasets. All relevant tools support CoNNL, including WebAnno for manual annotation. Nevertheless, it is possible, of course, to re-insert the annotated information back into the XML documents.
Court decisions from 2017 and 2018 were selected for the dataset, published online by the Federal Ministry of Justice and Consumer Protection.
|
What was selected for the dataset?
|
Court decisions from 2017 and 2018 were selected for the datase.
|
null | false
| null |
A convolutional neural network consists of an input layer, hidden layers and an output layer. In any feed-forward neural network, any middle layers are called hidden because their inputs and outputs are masked by the activation function and final convolution. In a convolutional neural network, the hidden layers include layers that perform convolutions. Typically this includes a layer that performs a dot product of the convolution kernel with the layer's input matrix. This product is usually the Frobenius inner product, and its activation function is commonly ReLU. As the convolution kernel slides along the input matrix for the layer, the convolution operation generates a feature map, which in turn contributes to the input of the next layer. This is followed by other layers such as pooling layers, fully connected layers, and normalization layers.
Convolutional layers
Convolutional layers convolve the input and pass its result to the next layer. This is similar to the response of a neuron in the visual cortex to a specific stimulus. Each convolutional neuron processes data only for its receptive field. Although fully connected feedforward neural networks can be used to learn features and classify data, this architecture is generally impractical for larger inputs (e.g., high-resolution images), which would require massive numbers of neurons because each pixel is a relevant input feature. A fully connected layer for an image of size 100 × 100 has 10,000 weights for each neuron in the second layer. Convolution reduces the number of free parameters, allowing the network to be deeper. For example, using a 5 × 5 tiling region, each with the same shared weights, requires only 25 neurons. Using regularized weights over fewer parameters avoids the vanishing gradients and exploding gradients problems seen during backpropagation in earlier neural networks.
To speed processing, standard convolutional layers can be replaced by depthwise separable convolutional layers, which are based on a depthwise convolution followed by a pointwise convolution. The depthwise convolution is a spatial convolution applied independently over each channel of the input tensor, while the pointwise convolution is a standard convolution restricted to the use of
Pooling layers
Convolutional networks may include local and/or global pooling layers along with traditional convolutional layers. Pooling layers reduce the dimensions of data by combining the outputs of neuron clusters at one layer into a single neuron in the next layer. Local pooling combines small clusters, tiling sizes such as 2 × 2 are commonly used. Global pooling acts on all the neurons of the feature map. There are two common types of pooling in popular use: max and average. Max pooling uses the maximum value of each local cluster of neurons in the feature map, while average pooling takes the average value.
Fully connected layers
Fully connected layers connect every neuron in one layer to every neuron in another layer. It is the same as a traditional multilayer perceptron neural network (MLP). The flattened matrix goes through a fully connected layer to classify the images.
Receptive field
In neural networks, each neuron receives input from some number of locations in the previous layer. In a convolutional layer, each neuron receives input from only a restricted area of the previous layer called the neuron's receptive field. Typically the area is a square (e.g. 5 by 5 neurons). Whereas, in a fully connected layer, the receptive field is the entire previous layer. Thus, in each convolutional layer, each neuron takes input from a larger area in the input than previous layers. This is due to applying the convolution over and over, which takes into account the value of a pixel, as well as its surrounding pixels. When using dilated layers, the number of pixels in the receptive field remains constant, but the field is more sparsely populated as its dimensions grow when combining the effect of several layers.
To manipulate the receptive field size as desired, there are some alternatives to the standard convolutional layer. For example, atrous or dilated convolution expands the receptive field size without increasing the number of parameters by interleaving visible and blind regions. Moreover, a single dilated convolutional layer can comprise filters with multiple dilation ratios, thus having a variable receptive field size.
Weights
Each neuron in a neural network computes an output value by applying a specific function to the input values received from the receptive field in the previous layer. The function that is applied to the input values is determined by a vector of weights and a bias (typically real numbers). Learning consists of iteratively adjusting these biases and weights.
The vectors of weights and biases are called filters and represent particular features of the input (e.g., a particular shape). A distinguishing feature of CNNs is that many neurons can share the same filter. This reduces the memory footprint because a single bias and a single vector of weights are used across all receptive fields that share that filter, as opposed to each receptive field having its own bias and vector weighting.
|
Please give me a short bulleted list of the key components/architectures in a convolutional neural network.
|
The key components/architecture of a convolutional neural network include:
Convolutional layers: These layers convolve the input matrix and reduce the number of free parameters and allows the network to be deeper.
Pooling layers: These layers reduce the dimensions of data by combining the outputs of neuron clusters at one layer into a single neuron in the next layer.
Fully connected layers: These layers connect every neuron in one layer to every neuron in another layer.
Receptive field: Each neuron receives input from a restricted area of the previous layer called the neuron's receptive field. In a convolutional layer, each neuron takes input from a larger area in the input than previous layers.
Weights: The vectors of weights and bias are used by the neural's function to compute an output value from input value.
|
null | false
| 145
|
All BERT experiments were trained for 500K steps. The code was written in Python and we used the Tensorflow libraries to create the models. The experiments were trained on a single NVIDIA Tesla V100 32 GB graphic card. The data was first processed into Tensorflow records as the input to the model. The set of hyperparameters which we found optimal after experimenting with different sets are in (Table TABREF10).
This set of parameters were chosen as there training performances were better than smaller models on modelling the long sequences of sub-words. We use the Adam optimizer BIBREF16 same as the English BERT. A maximum sequence length of 300 encompasses 98 percent of the training data and also allows us to fit larger models on the GPU card. Hyper-parameter optimization is very difficult in case of these models as they take around 15 days to train given the resources. The hyperparameter choices were therefore more dependant on the original BERT with little tweaks. We assess the training performance of the the model in the (Table TABREF11).
When we train the BERT model we mask some percentage of the input tokens at random, and then predict those masked tokens, this is known as Masked LM. The masked LM loss, refers specifically to the loss when the masked language model predicts on the masked tokens. The masked LM accuracy refers specifically to the accuracy with which the model predicts on the masked tokens. The loss for both the models are far off from the Masked LM loss of the English BERT, key difference being the pre-training data for both the language models are quite different. Google training their model on 3.3 Billion words from BooksCorpus BIBREF17 and the English Wikipedia and our model being trained on 144 million words. Comparing the two Finnish models, the left-marked model has a better training performance than left+right-marked model.
The results of the pseudo-perplexity described in the previous section to evaluate the above models on the test data-set is in table (Table TABREF12).The test dataset is of a different context when compared to the training data, and interestingly both the models are quite confident when it comes to the test dataset. The pseudo-perplexity values of left-marked are lower when compared to left-right-marked signifying that it is more confident.
We cannot directly compare the perplexity scores BERT model with a unidirectional LSTM model as both are calculated in a different manner. We can experiment to compare it with a Bi-directional LSTM or use a downstreaming task to compare both the performances. We could also randomly mask tokens and then compare the prediction accuracy on the masked tokens.
All Transformer-XL experiments are also trained equally for 500K steps. The code was written in Python and we used the PyTorch libraries for model creation. The experiments were trained on a single NVIDIA Tesla V100 32 GB graphic card. Two sets of hyperparameters were chosen to be compared after some initial optimization and are in (Table TABREF14)
From the above parameter choice, we wanted to experiment, whether providing more Segment and Memory length is advantageous (longer context) than a larger model. These parameters where chosen after some hyperparameter optimization. Same as for BERT we use the Adam optimizer, but we also use a cosine annealing learning rate scheduler to speed-up training BIBREF18. The training performance results are in (Table TABREF15)
As opposed to BERT, the left+right-marked models have a better training performance than their counterpart. Interestingly the larger model trains much better when compared to providing larger contexts. The same set of parameters for the 32-32 model cannot be replicated for 150-150 model as the latter takes a lot of space on the GPU card. The test set is same as that used with BERT and the results are in (Table TABREF16). The test performance is similar to that of the training performance with left-right-marked large model(32-32) performing the best. We can directly compare the perplexity scores with the previous best BIBREF19 as both are unidirectional models, Transformer-XL model has outperformed the latter by 27%.
Transformer-XL and BERT both have low perplexity and pseudo-perplexity scores, but both cannot be directly compared as they are calculated quite differently (Eq.DISPLAY_FORM4, Eq.DISPLAY_FORM6). The dramatically low scores of BERT indicate that per word predicted probability is higher than that of a uni-directional model. Thus the predicted word probability distribution is much sharper when compared to the XL model probability distribution. At this point, we cannot say which model architecture has performed better- BERT or Transformer-XL, despite both of them achieving good low perplexity scores. We would need to experiment with a downstreaming task in-order to fairly compare model performances.
Transformer-XL and BERT both have low perplexity and pseudo-perplexity scores, but both cannot be directly compared as they are calculated quite differently (Eq.1, Eq.3).
|
Whether Transformer-XL and BERT both have low perplexity and pseudo-perplexity scores?
|
Yes.
|
null | false
| null |
In passive solar building design, windows, walls, and floors are made to collect, store, reflect, and distribute solar energy, in the form of heat in the winter and reject solar heat in the summer. This is called passive solar design because, unlike active solar heating systems, it does not involve the use of mechanical and electrical devices.
The key to designing a passive solar building is to best take advantage of the local climate performing an accurate site analysis. Elements to be considered include window placement and size, and glazing type, thermal insulation, thermal mass, and shading. Passive solar design techniques can be applied most easily to new buildings, but existing buildings can be adapted or "retrofitted".
|
What is passive solar building design
|
In passive solar building design, windows, walls, and floors are made to collect, store, reflect, and distribute solar energy, in the form of heat in the winter and reject solar heat in the summer. This is called passive solar design because, unlike active solar heating systems, it does not involve the use of mechanical and electrical devices.
The key to designing a passive solar building is to best take advantage of the local climate performing an accurate site analysis. Elements to be considered include window placement and size, and glazing type, thermal insulation, thermal mass, and shading. Passive solar design techniques can be applied most easily to new buildings, but existing buildings can be adapted or "retrofitted".
|
null | false
| 59
|
Conventional automatic speech recognition (ASR) systems typically consist of several independently learned components: an acoustic model to predict context-dependent sub-phoneme states (senones) from audio, a graph structure to map senones to phonemes, and a pronunciation model to map phonemes to words. Hybrid systems combine hidden Markov models to model state dependencies with neural networks to predict states BIBREF0 , BIBREF1 , BIBREF2 , BIBREF3 . Newer approaches such as end-to-end (E2E) systems reduce the overall complexity of the final system.
Our research builds on prior work that has explored using time-delay neural networks (TDNN), other forms of convolutional neural networks, and Connectionist Temporal Classification (CTC) loss BIBREF4 , BIBREF5 , BIBREF6 . We took inspiration from wav2letter BIBREF6 , which uses 1D-convolution layers. Liptchinsky et al. BIBREF7 improved wav2letter by increasing the model depth to 19 convolutional layers and adding Gated Linear Units (GLU) BIBREF8 , weight normalization BIBREF9 and dropout.
By building a deeper and larger capacity network, we aim to demonstrate that we can match or outperform non end-to-end models on the LibriSpeech and 2000hr Fisher+Switchboard tasks. Like wav2letter, our architecture, Jasper, uses a stack of 1D-convolution layers, but with ReLU and batch normalization BIBREF10 . We find that ReLU and batch normalization outperform other activation and normalization schemes that we tested for convolutional ASR. As a result, Jasper's architecture contains only 1D convolution, batch normalization, ReLU, and dropout layers – operators highly optimized for training and inference on GPUs.
It is possible to increase the capacity of the Jasper model by stacking these operations. Our largest version uses 54 convolutional layers (333M parameters), while our small model uses 34 (201M parameters). We use residual connections to enable this level of depth. We investigate a number of residual options and propose a new residual connection topology we call Dense Residual (DR).
Integrating our best acoustic model with a Transformer-XL BIBREF11 language model allows us to obtain new state-of-the-art (SOTA) results on LibriSpeech BIBREF12 test-clean of 2.95% WER and SOTA results among end-to-end models on LibriSpeech test-other. We show competitive results on Wall Street Journal (WSJ), and 2000hr Fisher+Switchboard (F+S). Using only greedy decoding without a language model we achieve 3.86% WER on LibriSpeech test-clean.
This paper makes the following contributions:
We introduce NovoGrad, a variant of the Adam optimizer [15] with a smaller memory footprint.
|
What optimizer does the team introduce in their research?
|
NovoGrad.
|
null | false
| null |
Classify the below Noble prize winners based on the country they are from.
Rabindranath Tagore, Chandrasekhara Venkata Raman, Amartya Kumar Sen, Joseph John Thomson, Ernest Rutherford, Niels Bohr
|
India- Rabindranath Tagore, Chandrasekhara Venkata Raman, Amartya Kumar Sen, Joseph John Thomson, Ernest Rutherford, Niels Bohr
UK- Joseph John Thomson
New Zealand - Ernest Rutherford
Denmark - Niels Bohr
|
|
null | false
| null |
Allen Eager was born in New York City on January 10, 1927. He grew up in the Bronx. According to Denise McCluggage, Eager could read aged 3, and learned to drive at the age of 9 with the help of his mother, after she caught him driving a garbage truck near the hotels that his parents owned in the Catskill Mountains. He took clarinet lessons with David Weber of the New York Philharmonic at the age of 13.
|
Where did Allen Eager grow up?
|
Allen Eager grew up in the Bronx.
|
null | false
| null |
A glass ceiling is a metaphor usually applied to women, used to represent an invisible barrier that prevents a given demographic from rising beyond a certain level in a hierarchy. No matter how invisible the glass ceiling is expressed, it is actually a difficult obstacle to overcome. The metaphor was first used by feminists in reference to barriers in the careers of high-achieving women. It was coined by Marilyn Loden during a speech in 1978. In the United States, the concept is sometimes extended to refer to racial inequality in the United States. Minority women in white-majority countries often find the most difficulty in "breaking the glass ceiling" because they lie at the intersection of two historically marginalized groups: women and people of color. East Asian and East Asian American news outlets have coined the term "bamboo ceiling" to refer to the obstacles that all East Asian Americans face in advancing their careers. Similarly, a multitude of barriers that refugees and asylum seekers face in their search for meaningful employment is referred to as canvas ceiling.
|
Why is the glass ceiling better for minority women?
|
Actually, the glass ceiling is considered worse for minority women, not better. This is because they have to face both racial and gender inequalities.
|
null | false
| null |
What is the mascot of Cornell University?
|
Cornell's mascot is "Big Red" or "Touchdown the Bear." It's common to hear Cornellian's chant "Go Big Red" at sporting events.
|
|
null | false
| null |
Which state's capital is inaccessible by road?
|
Juneau, Alaska is inaccessible by road, but cars can get to the city via ferry.
|
|
2002.01359
| false
| null |
We consider the following metrics for automatic evaluation of different submissions. Joint goal accuracy has been used as the primary metric to rank the submissions.
Active Intent Accuracy: The fraction of user turns for which the active intent has been correctly predicted.
Requested Slot F1: The macro-averaged F1 score for requested slots over all eligible turns. Turns with no requested slots in ground truth and predictions are skipped.
Average Goal Accuracy: For each turn, we predict a single value for each slot present in the dialogue state. This is the average accuracy of predicting the value of a slot correctly.
Joint Goal Accuracy: This is the average accuracy of predicting all slot assignments for a given service in a turn correctly.
We consider the following metrics for automatic evaluation of different submissions. Joint goal accuracy has been used as the primary metric to rank the submissions.
Active Intent Accuracy: The fraction of user turns for which the active intent has been correctly predicted.
Requested Slot F1: The macro-averaged F1 score for requested slots over all eligible turns. Turns with no requested slots in ground truth and predictions are skipped.
Average Goal Accuracy: For each turn, we predict a single value for each slot present in the dialogue state. This is the average accuracy of predicting the value of a slot correctly.
Joint Goal Accuracy: This is the average accuracy of predicting all slot assignments for a given service in a turn correctly.
|
How are the models evaluated?
|
The answers are shown as follows:
* Active Intent Accuracy
* Requested Slot F1
* Average Goal Accuracy
* Joint Goal Accuracy
|
null | false
| null |
Can AI take my job
|
While it is possible for AI to do a piece of work that is done by a human, it is unlikely that AI will make humans redundant in the workplace. What is more plausible is that AI will change the nature of how humans work, enabling greater productivity or becoming a validation/verification process.
A good example of this is GPS. For years, taxi drivers used their knowledge of an area to take people from one place to another, but with the invention of GPS, this local knw;edge became less valuable (but not redundant). There was not a reduction in Taxi drives, if anything, there were more Taxis available to the general public.
The spell checker has been in software for decades, but it is still down to the individual to clarify their specific meaning and validate the final output, rather than relying purely on what a computer generates.
|
|
null | false
| null |
What are the three major Abrahamic religions?
|
Judaism, Christianity and Islam are the three largest Abrahamic religions
|
|
null | false
| null |
List 10 best switch games of all time.
|
According to the metacritic website, the top 10 switch games of all time are:
1. The Legend of Zelda: Breath of the Wild
2. Super Mario Odyssey
3. Portal Companion Collection
4. The House in Fata Morgana - Dreams of the Revenants Edition
5. Tetris Effect: Connected
6. Metroid Prime Remastered
7. Persona 5 Royal
8. Hades
9. Divinity: Original Sin II - Definitive Edition
10. Ori and the Will of the Wisps
The top 10 lists from other websites may vary.
|
|
null | false
| null |
Best places to visit in India During Winter to experience Snowfall
|
Kashmir, Manali, Shimla, Sikkim, Tawang
|
|
null | false
| null |
Todd, third son of Ira and Sally (Hmman) Todd, was born in Hartwick, New York., March 4, 1813. He had spent one year at Amherst College, before entering as Sophomore at Yale College, where he graduated in 1836. He taught a high school in Canaan, Conn., until the spring of 1837, when he began law studies with Judge Arphaxad Loomis, of Little Falls, New York. Upon his admission to the bar in 1839, he decided to emigrate to the West. He selected St. Louis and arrived there in November, 1839, beginning practice in March, 1840. He gave special attention to questions affecting real property, and achieved distinguished success in litigated cases of this nature. Through apprehensions with regard to his health, he retired in 1860, at a comparatively early age, from active practice, confining himself thereafter to office consultations. He declined all political nominations until 1854, when he was elected to the Missouri House of Representatives. In 1860 he was an unsuccessful candidate for US Congress on the Bell and Everett ticket. In 1875, he was a member of the convention which revised the Constitution of Missouri. He was a director of Washington University, in St. Louis, and served gratuitously for fifteen years in its Law Department as lecturer. Among his students was Lemma Barkaloo, the first female law student in the United States. After a month's illness, he died at his home in St. Louis, of meningitis, April 30, 1885, in his 73rd year.
|
Extract all the locations in the text
|
Hartwick, New York
Canaan, Connecticut
Little Falls, New York
St Louis
|
null | false
| null |
What is it like to live in New Zealand?
|
New Zealand is considered a really safe country to live in. Politics are quite liberal with the left and the right political parties not having a lot of difference between them. The climate is very diverse from the top of the North Island to the bottom of the South Island however all seasons are generally pretty mild without many extreme weather events. Housing is a big issue in New Zealand, with house prices being extremely high compared to income earned. The culture is could be described as laid back and easy going.
|
|
null | false
| 247
|
With the rapid growth of social network platforms, more and more people tend to share their experiences and emotions online.[2]Corresponding Author: xuruifeng@hit.edu.cn Emotion analysis of online text becomes a new challenge in Natural Language Processing (NLP). In recent years, studies in emotion analysis largely focus on emotion classification including detection of writers' emotions BIBREF0 as well as readers' emotions BIBREF1 . There are also some information extraction tasks defined in emotion analysis BIBREF2 , BIBREF3 , such as extracting the feeler of an emotion BIBREF4 . These methods assume that emotion expressions are already observed. Sometimes, however, we care more about the stimuli, or the cause of an emotion. For instance, Samsung wants to know why people love or hate Note 7 rather than the distribution of different emotions.
Ex.1 我的手机昨天丢了，我现在很难过。
Ex.1 Because I lost my phone yesterday, I feel sad now.
In an example shown above, “sad” is an emotion word, and the cause of “sad” is “I lost my phone”. The emotion cause extraction task aims to identify the reason behind an emotion expression. It is a more difficult task compared to emotion classification since it requires a deep understanding of the text that conveys an emotions.
Existing approaches to emotion cause extraction mostly rely on methods typically used in information extraction, such as rule based template matching, sequence labeling and classification based methods. Most of them use linguistic rules or lexicon features, but do not consider the semantic information and ignore the relation between the emotion word and emotion cause. In this paper, we present a new method for emotion cause extraction. We consider emotion cause extraction as a question answering (QA) task. Given a text containing the description of an event which [id=lq]may or may not cause a certain emotion, we take [id=lq]an emotion word [id=lq]in context, such as “sad”, as a query. The question to the QA system is: “Does the described event cause the emotion of sadness?”. The [id=lq]expected answer [id=lq]is either “yes” or “no”. (see Figure FIGREF1 ). We build our QA system based on a deep memory network. The memory network has two inputs: a piece of text, [id=lq]referred to as a story in QA systems, and a query. The [id=lq]story is represented using a sequence of word embeddings.
[id=lq]A recurrent structure is implemented to mine the deep relation between a query and a text. It measure[id=lq]s the [id=lq]importance of each word in the text by [id=lq]an attention mechanism. Based on the [id=lq]learned attention result, the network maps the text into a low dimensional vector space. This vector is [id=lq]then used to generate an answer. Existing memory network based approaches to QA use weighted sum of attentions to jointly consider short text segments stored in memory. However, they do not explicitly model [id=lq]sequential information in the context. In this paper, we propose a new deep memory network architecture to model the context of each word simultaneously by multiple memory slots which capture sequential information using convolutional operations BIBREF5 , and achieves the state-of-the-art performance compared to existing methods which use manual rules, common sense knowledge bases or other machine learning models.
The rest of the paper is organized as follows. Section SECREF2 gives a review of related works on emotion analysis. Section SECREF3 presents our proposed deep memory network based model for emotion cause extraction. Section SECREF4 discusses evaluation results. Finally, Section SECREF5 concludes the work and outlines the future directions.
In this paper, we propose a new deep memory network architecture to model the context of each word simultaneously by multiple memory slots which capture sequential information using convolutional operations (Kim, 2014), and achieves the state-of-the-art performance compared to existing methods which use manual rules, common sense knowledge bases or other machine learning models.
|
Does the new model achieve the state-of-the-art performance compared to existing methods?
|
Yes, it does.
|
null | false
| null |
Which TV show features a team of teenagers that power up with powers given to them by Zordon to save the world from monsters?
|
The Mighty Morphin Power Rangers.
|
|
null | false
| null |
What are some of the most popular tourist attractions to visit in Sydney?
|
Sydney has many famous landmarks and attractions, with the most famous being the Sydney Harbour Bridge and the Sydney Opera House, situated on the foreshore of Sydney harbour. These two landmarks are the iconic images of Sydney and are often viewed from public ferries, either a ferry taken to visit Manly Beach or to Watsons bay. Visitors to Sydney also often visit the historical Rocks district to see the oldest surviving buildings in the city and learn about the town's early history. The view from the top of the Sydney Tower is also popular, with views out to the Pacific Ocean over the famous Bondi Beach. Also located on Sydney harbour is the Taronga Park Zoo, with animals both domestic and imported and a spectacular view over the harbour. Further out of town it is common to visit the Blue Mountains, with the Three Sisters rock formation at Katoomba being particularly popular.
|
|
null | false
| null |
Describe a scary moment in a storm.
|
I live in Oklahoma, so we see our fair share of storms. Storm sirens are so common that I don't take cover every time they go off. Instead, it's normal for me to watch the radar when a storm system is moving through to see where the tornados are and how close they are to me.
Once, the radar showed a tornado southwest of our house - moving towards us. When it was within a couple miles of us, my wife took the dogs and blankets and went to our safe place - an internal bathroom. I stood back from the windows and watched the tv. From my position, I could see out the back windows which faced west - southwest. The weather showed that the tornado was right on top of us. I watched out the window as a fragment of a four by four, about a foot and a half long, flipped up from the ground and tumbled across the lawn.
It was that moment that I decided I had watched enough and stepped into the bathroom to join my wife and the dogs. The storm blew over us. The tornado wasn't on the ground as it passed by, but it was definitely and intense several minutes as we waited.
I have seen my fair share of storms and had a number of close calls with tornados. This definitely falls in one of my top 5.
|
|
null | false
| 144
|
Multi-document summarization (MDS), the transformation of a set of documents into a short text containing their most important aspects, is a long-studied problem in NLP. Generated summaries have been shown to support humans dealing with large document collections in information seeking tasks BIBREF0 , BIBREF1 , BIBREF2 . However, when exploring a set of documents manually, humans rarely write a fully-formulated summary for themselves. Instead, user studies BIBREF3 , BIBREF4 show that they note down important keywords and phrases, try to identify relationships between them and organize them accordingly. Therefore, we believe that the study of summarization with similarly structured outputs is an important extension of the traditional task.
A representation that is more in line with observed user behavior is a concept map BIBREF5 , a labeled graph showing concepts as nodes and relationships between them as edges (Figure FIGREF2 ). Introduced in 1972 as a teaching tool BIBREF6 , concept maps have found many applications in education BIBREF7 , BIBREF8 , for writing assistance BIBREF9 or to structure information repositories BIBREF10 , BIBREF11 . For summarization, concept maps make it possible to represent a summary concisely and clearly reveal relationships. Moreover, we see a second interesting use case that goes beyond the capabilities of textual summaries: When concepts and relations are linked to corresponding locations in the documents they have been extracted from, the graph can be used to navigate in a document collection, similar to a table of contents. An implementation of this idea has been recently described by BIBREF12 .
The corresponding task that we propose is concept-map-based MDS, the summarization of a document cluster in the form of a concept map. In order to develop and evaluate methods for the task, gold-standard corpora are necessary, but no suitable corpus is available. The manual creation of such a dataset is very time-consuming, as the annotation includes many subtasks. In particular, an annotator would need to manually identify all concepts in the documents, while only a few of them will eventually end up in the summary.
To overcome these issues, we present a corpus creation method that effectively combines automatic preprocessing, scalable crowdsourcing and high-quality expert annotations. Using it, we can avoid the high effort for single annotators, allowing us to scale to document clusters that are 15 times larger than in traditional summarization corpora. We created a new corpus of 30 topics, each with around 40 source documents on educational topics and a summarizing concept map that is the consensus of many crowdworkers (see Figure FIGREF3 ).
As a crucial step of the corpus creation, we developed a new crowdsourcing scheme called low-context importance annotation. In contrast to traditional approaches, it allows us to determine important elements in a document cluster without requiring annotators to read all documents, making it feasible to crowdsource the task and overcome quality issues observed in previous work BIBREF13 . We show that the approach creates reliable data for our focused summarization scenario and, when tested on traditional summarization corpora, creates annotations that are similar to those obtained by earlier efforts.
To summarize, we make the following contributions: (1) We propose a novel task, concept-map-based MDS (§ SECREF2 ), (2) present a new crowdsourcing scheme to create reference summaries (§ SECREF4 ), (3) publish a new dataset for the proposed task (§ SECREF5 ) and (4) provide an evaluation protocol and baseline (§ SECREF7 ). We make these resources publicly available under a permissive license.
To summarize, we make the following contributions: (1) We propose a novel task, concept-mapbased MDS (§2), (2) present a new crowdsourcing scheme to create reference summaries (§4), (3) publish a new dataset for the proposed task (§5) and (4) provide an evaluation protocol and baseline (§7). We make these resources publicly available under a permissive license.
|
What are their contributions?
|
They propose concept-mapbased MDS which is a totally novel task and create a new dataset and a new crowdsourcing scheme for it, they also provide an evaluation protocol and baseline.
|
null | false
| null |
Write a short passage explaining why someone should become a fan of the Pittsburgh Steelers.
|
The Pittsburgh Steelers are one of the great franchises in the National Football League. The Steelers have a rich and storied history, including six Super Bowl victories. The Steelers have been owned by the Rooney family from their inception, and are renowned for their organizational excellence. The Steelers are also known for having a vast global diaspora of fans, all of whom proudly wave and display their "Terrible Towels" to cheer on them on. Additionally, their current roster is filled with exciting young talent, including Kenny Pickett, George Pickens, and Najee Harris. The Steelers are also leaders in propagating diversity within their organization, especially within their leadership. Lastly, Mike Tomlin, the head coach of the Pittsburgh Steelers, is a proven winner, and has never had a losing season while coaching the Steelers. The Steelers are a fun team to watch, a fun team to cheer for, and a team that has great pride in their values on and off the field.
|
|
null | false
| 310
|
Quality estimation (QE) refers to the task of measuring the quality of machine translation (MT) system outputs without reference to the gold translations BIBREF0 , BIBREF1 . QE research has grown increasingly popular due to the improved quality of MT systems, and potential for reductions in post-editing time and the corresponding savings in labor costs BIBREF2 , BIBREF3 . QE can be performed on multiple granularities, including at word level, sentence level, or document level. In this paper, we focus on quality estimation at word level, which is framed as the task of performing binary classification of translated tokens, assigning “OK” or “BAD” labels.
Early work on this problem mainly focused on hand-crafted features with simple regression/classification models BIBREF4 , BIBREF5 . Recent papers have demonstrated that utilizing recurrent neural networks (RNN) can result in large gains in QE performance BIBREF6 . However, these approaches encode the context of the target word by merely concatenating its left and right context words, giving them limited ability to control the interaction between the local context and the target word.
In this paper, we propose a neural architecture, Context Encoding Quality Estimation (CEQE), for better encoding of context in word-level QE. Specifically, we leverage the power of both (1) convolution modules that automatically learn local patterns of surrounding words, and (2) hand-crafted features that allow the model to make more robust predictions in the face of a paucity of labeled data. Moreover, we further utilize stacked recurrent neural networks to capture the long-term dependencies and global context information from the whole sentence.
We tested our model on the official benchmark of the WMT18 word-level QE task. On this task, it achieved highly competitive results, with the best performance over other competitors on English-Czech, English-Latvian (NMT) and English-Latvian (SMT) word-level QE task, and ranking second place on English-German (NMT) and German-English word-level QE task.
We tested our model on the official benchmark of the WMT18 word-level QE task.
|
What task is set for testing the model?
|
The WMT18 word-level QE task.
|
null | false
| 63
|
Rhetorical Structure Theory (RST) BIBREF0 is a technique of Natural Language Processing (NLP), in which a document can be structured hierarchically according to its discourse. The generated hierarchy, a tree, provides information associated with the boundaries of the discourse segments and related to their importance and dependencies. The figure FIGREF1 shows an example of such a rethorical tree. In the rethorical parsing process, the text has been divided into five units. In the figure FIGREF1, the arrow that leaves the unit (2) towards the unit (1) symbolizes that the unit (2) is the satellite of the unit (1), which is the core in a “Concession” relationship. In turn, the units (1) and (2) comprise the nucleus of three “Demonstration” relationships.
The discursive analysis of a document normally includes three consecutive steps: 1) discursive segmentation; 2) detection of the discursive relations; 3) construction of the hierarchical rhetorical tree. Regarding the discursive segmentation, there are segmenters in several languages. However, each piece depends on sofisticated linguistic resources, which complicates the reproduction of the experiments in other languages. Consequently, the development of multilingual systems using discursive analysis are yet to be developed. Diverse applications based on the latest technologies require at least one of the three steps mentioned above BIBREF1, BIBREF2, BIBREF3. In this context, the idea of exploring the architecture of a generic system that is able not only of segmenting a text correctly but also of adapting it to any language, was a great motivation of this research work.
In this article we show the preliminary results of a generic segmenter composed of several systems (different segmentation strategies). In addition, we describe an automatic evaluation protocol of discursive segmentation. The article is composed by the following sections: state of the art (SECREF2), which presents a brief bibliographic review; Description of the Annodis (SECREF3) corpus used in our tests and of the general architecture of the proposed systems (SECREF4); Segmentation strategies (SECREF5), which characterizes the different methods implemented to segment the text; results of our numerical experiments (sec:experiments); and we conclude with our conclusions and perspectives (SECREF7).
In the figure 1, the arrow that leaves the unit (2) towards the unit (1) symbolizes that the unit (2) is the satellite of the unit (1), which is the core in a “Concession” relationship. In turn, the units (1) and (2) comprise the nucleus of three “Demonstration” relationships. The discursive analysis of a document normally includes three consecutive steps: 1) discursive segmentation; 2) detection of the discursive relations; 3) construction of the hierarchical rhetorical tree.
|
How many steps does discursive analysis of a document usually include?
|
Three consecutive steps. 1) discursive segmentation; 2) detection of the discursive relations; 3) construction of the hierarchical rhetorical tree.
|
null | false
| null |
What are people from Wisconsin also referred to as?
|
Some people refer to people from Wisconsin as Cheeseheads because of the amount of dairy that is produced by the state. It was a derogatory term set in from a local sports rivalry with people from Illinois, however the Cheeseheads from Wisconsin embraced it, and now wear Cheesehead hats to local sporting events.
|
|
1911.08829
| true
| null |
For parser-based extraction, systems with and without in-context parsing, ignoring labels, and ignoring directionality are tested. For the three string-based extraction methods, varying numbers of intervening words and case sensitivity are evaluated. Evaluation is done using the development set, consisting of 22 documents and 1112 PIE candidates, and the test set, which consists of 23 documents and 1127 PIE candidates. For each method the best set of parameters and/or options is determined using the development set, after which the best variant by F1-score of each method is evaluated on the test set.
Since these documents in the corpus are exhaustively annotated for PIEs (see Section SECREF40), we can calculate true and false positives, and false negatives, and thus precision, recall and F1-score. The exact spans are ignored, because the spans annotated in the evaluation corpus are not completely reliable. These were automatically generated during candidate extraction, as described in Section SECREF45. Rather, we count an extraction as a true positive if it finds the correct PIE type in the correct sentence.
Evaluation is done using the development set, consisting of 22 documents and 1112 PIE candidates, and the test set, which consists of 23 documents and 1127 PIE candidates. For each method the best set of parameters and/or options is determined using the development set, after which the best variant by F1-score of each method is evaluated on the test set.
Since these documents in the corpus are exhaustively annotated for PIEs (see Section SECREF40), we can calculate true and false positives, and false negatives, and thus precision, recall and F1-score.
|
Are PIEs extracted automatically subjected to human evaluation?
|
Yes.
|
null | false
| null |
Which of the following from this list are types of flowers and which are names of European countries: roses, Norway, tulips, the Netherlands, Sweden, France, Spain, Greece, Italy, and sunflowers.
|
The types of flowers from this list are: tulips, roses and sunflowers. The European countries are Norway, the Netherlands, Sweden, France, Spain, Greece, and Italy.
|
|
null | false
| null |
What is the difference between a guitar and a ukulele?
|
While a ukulele might be thought of as a small guitar, there are actually numerous differences between the two instruments:
1. A ukulele has 4 strings while a guitar has 6 strings or sometimes 12
2. A ukulele's 4 strings are tuned differently from any consecutive series of a guitar's 4 strings
|
|
null | false
| null |
Outcomes from Project Vesta have been integrated into firefighter training in Australia and are beginning to appear in the United States.
Firefighters try to stay out of the dead man zone at all times, working from safe points such as burnt ground or a large area of non-burnable ground, such as a cricket or Australian rules football oval, or a large car park. This is achieved by attacking the fire from the flanks, or the rear, so that burnt ground is always nearby, and the fire is always in front of the firefighters. This avoids two disadvantages of attacking fires at the head of the fire where spot fires may start behind them or changes in wind behavior might accelerate the spread of the fire.
The result of several inquiries into firefighter death in Australian bushfires found that firefighters should stay out of the dead man zone and that they should always keep 250 liters of water in their truck for personal safety. This is now a standard operating procedure in the NSW Rural Fire Service, Country Fire Service and Country Fire Authority in Australia.
|
What are some areas that firefighters can safely stay in while fighting a wildfire?
|
Firefighters can use burnt areas, car parks and sports areas to stay safe during a wildfire
|
null | false
| null |
Why do people get married?
|
People get married for a variety of reasons. Most people get married because they feel like they are in love with the other person and want to spend the rest of their life with that person. Many people get married because they want to start a family. Some get married in order to have tax advantages and legal protection.
|
|
1709.08858
| false
| null |
Distributed representation of word sense provides us with the ability to perform several operations on the word. One of the most important operations on a word is to obtain the set of words whose meaning is similar to the word, or whose usage in text is similar to the word. We call this set the neighbor of the word. When a word has several senses, it is called a polysemic word. When a word has only one sense, it is called a monosemic word. We have observed that the neighbor of a polysemic word consists of words that resemble the primary sense of the polysemic word. We can explain this fact as follows. Even though a word may be a polysemic, it usually corresponds to a single vector in distributed representation. This vector is primarily determined by the major sense, which is most frequently used. The information about a word's minor sense is subtle, and the effect of a minor sense is difficult to distinguish from statistical fluctuation.
To measure the effect of a minor sense, this paper proposes to use the concept of surrounding uniformity. The surrounding uniformity roughly corresponds to statistical fluctuation in the vectors that correspond to the words in the neighbor. We have found that there is a difference in the surrounding uniformity between a monosemic word and a polysemic word. This paper describes how to compute surrounding uniformity for a given word, and discuss the relationship between surrounding uniformity and polysemy.
We choose the uniformity of vectors, which can be regarded as general case of triangle inequality. The uniformity of a set of vectors is a ratio, i.e., the size of the vector of the vector addition of the vectors divided by the scalar sum of the sizes of the vectors. If and only if all directions of the vectors are the same, the uniformity becomes 1.0. We compute this uniformity for the neighbors, including the word itself. Surrounding Uniformity (SU) can be expressed as follows: $SU(\vec{w}) = \frac{|\vec{s}(\vec{w})|}{|\vec{w}| + \sum _{i}^{N}|\vec{a_i}(\vec{w})|}$
where $\vec{s}(\vec{w}) = \vec{w} + \sum _{i}^{N} \vec{a_i}(\vec{w}).$
One of the most important operations on a word is to obtain the set of words whose meaning is similar to the word, or whose usage in text is similar to the word. We call this set the neighbor of the word.
We have observed that the neighbor of a polysemic word consists of words that resemble the primary sense of the polysemic word.
Even though a word may be a polysemic, it usually corresponds to a single vector in distributed representation. This vector is primarily determined by the major sense, which is most frequently used. The information about a word's minor sense is subtle, and the effect of a minor sense is difficult to distinguish from statistical fluctuation.
To measure the effect of a minor sense, this paper proposes to use the concept of surrounding uniformity.
The surrounding uniformity roughly corresponds to statistical fluctuation in the vectors that correspond to the words in the neighbor
Surrounding Uniformity (SU) can be expressed as follows: $SU(\vec{w}) = \frac{|\vec{s}(\vec{w})|}{|\vec{w}| + \sum _{i}^{N}|\vec{a_i}(\vec{w})|}$
where $\vec{s}(\vec{w}) = \vec{w} + \sum _{i}^{N} \vec{a_i}(\vec{w}).$
|
How is the fluctuation in the sense of the word and its neighbors measured?
|
Our method performs a statistical test to determine whether a given word is used polysemously in the text, according to the following steps:
1) Setting N, the size of the neighbor.
2) Choosing N neighboring words ai in the order whose angle with the vector of the given word w is the smallest.
3) Computing the surrounding uniformity for ai(0 < i ≤ N) and w.
4) Computing the mean m and the sample variance σ for the uniformities of ai .
5) Checking whether the uniformity of w is less than m − 3σ. If the value is less than m − 3σ, we may regard w as a polysemic word.
|
null | false
| null |
What year was the car manufacturer Ferrari founded in?
|
Ferrari was founded in 1939 from the Alfa Romeo racing division as Auto Avio Costruzioni, the company built its first car in 1940, and produced its first Ferrari-badged car in 1947.
|
|
null | false
| null |
What is the biggest land mammal?
|
The answer depends on how you measure size and whether you are talking about living animals or extinct animals.
The heaviest living land mammal is the African elephant at 7 tons. The tallest is the giraffe at 19 feet.
The largest extinct land mammal is not known for sure, and new fossils may be discovered at any time, but in 2021 palaeontologists claimed that a new species of giant rhino - Paraceratherium linxiaense - was likely the heaviest known at around 24 tons.
|
|
null | false
| null |
How do we increase voter turnout in the United States?
|
Voter turnout in the United States varies from state to state and, as a whole, is comparatively low with respect to other democracies. There are institutional as well as psychological factors that affect turnout. Institutional factors include voter registration requirements, timing and frequency of elections, location of polling places, and complexity of ballots. Psychological factors include political efficacy, interest and engagement of the electorate, and political party membership. Voter turnout in the United States can be increased, and I will propose two simple reforms that would increase voter turnout. I will also propose another more complex reform that would dramatically increase voter turnout but would be difficult to implement by the state governments.
Several institutional factors determine voter turnout. As each state has the power to hold elections, these factors vary across the states. The difficulty or ease of voter registration will have an impact on turnout. Minnesota has a very high turnout allowing same-day voter registration, while Texas has a low turnout and has a 30-day requirement. The day and time an election is also held matters. For example, the U.S. holds federal elections on a Tuesday in November during work hours which reduces voter turnout. The frequency of elections will have an influence on turnout. There are many elections per year in the U.S., and voters get fatigued and stop going to the polls. Voters are also less likely to vote if they don't have convenient polling places that are nearby and easy to access. Schools are the best and most accessible polling locations, but the U.S. holds elections on Tuesdays, which limits the availability of school space and parking. Lastly, complex ballots diminish enthusiasm and negatively impact U.S. voter turnout.
Along with institutional factors, there are psychological factors that control voter turnout. One of which is political efficacy which is defined in two ways. The first is internal efficacy which is how well you think you are able to understand politics. The second is external efficacy which is how well you feel the system responds to your input. Essentially, if you think you are smart enough and your vote "matters," you have high efficacy and tend to go to the polls. The next psychological factor is interest, which is a measure of how much you care about politics. If you don't care, it's highly unlikely that you will vote. The last psychological factor is partisanship. Are you a member of a political party? If yes, then it's more likely you'll go vote to support your party.
By voting age population, turnout in the U.S. is around 53%. This shows the United States has a low turnout compared to other democracies and puts the U.S. 7th from the bottom when compared to 35 other democracies. Australia has the highest turnout, with just over 90%. The reason for Australia's high turnout is that voting is compulsory, with fines for people who don't vote. "Voting in Australia is like a party," with election day described as a country-wide BBQ with easy-to-access polling locations held on a Saturday. In Australia, "Forcing people to engage in the process increases their knowledge of the issues and candidates," thereby increasing Australia's political interest and engagement. By contrast, polls in the U.S. consistently show us that Americans don't know much about politics and are neither engaged nor interested in politics. Converse argues that people in the U.S. have low levels of ideological constraint and conceptualization. This shows that Americans have low political efficacy.
In general American turnout is low, but Voter Eligible Population (VEP) turnout varies dramatically across the states. Hawaii has the least VEP turnout, with approximately 43%. To compare, the state with the highest VEP turnout is Minnesota, with 75%. Texas comes in 3rd from the bottom with 53%. Minnesota's high turnout is explained by "easy access to the ballot," "a sense of civic responsibility," "high rates of educational attainment and income," and "competitive and interesting elections." By comparison, the main reasons Hawaii turnout is so low are that it's "hard to register," voter disinterest and low efficacy, and it's a "one-party state". This combination of institutional and psychological factors, beneficial in Minnesota, and detrimental in Hawaii, has a significant impact on a state's turnout.
I propose two simple solutions to increase voter turnout. The first would be to move elections to the weekend or require that companies give employees the day off to vote. Sunday is the most common election day in the world outside the United States. I believe this would be a beneficial institutional factor that would increase voter turnout because the majority of other countries have higher voter turnout than the United States. Unfortunately, the chances of this happening are slim as Congress has proposed bills to move elections to the weekend or have a Federal holiday for voting, but these bills have never moved out of committees.
The second simple solution would address the institutional factor of registration requirements by implementing same-day voter registration in all 50 states. Minnesota allows for same-day registration, and its turnout is around 20% higher than Texas, which does not allow it. States that have implemented same-day voter registration have seen an average of a 5% increase in voters ("Same Day Voter"). To prevent voter fraud, systems should be put in place to make sure the voter hasn't already voted and to check for valid identification showing the voter is a resident of the state or county. Another deterrent to committing voter fraud would be criminal penalties, including jail time and fines. Enacting legislation in each state to implement a system to handle would take time, but I think it's possible as currently, 21 states allow same-day voting registration ("Same Day Voter").
The third solution I propose would address most of the institutional factors. This would be an online/app-based voting option. The vast majority of Americans these days have mobile phones with internet connections. Those that don't would still have all of the current voting options, such as polling locations and mail-in voting. Security and verification would combine blockchain (similar to Bitcoin identities), biometrics, and facial recognition artificial intelligence. When registering to vote, a state agency would create your blockchain voter identification, fingerprint, and take a LiDAR(Light Detection and Ranging) 3D map of your face. These three technologies would ensure one vote per person. Also, this technology is all readily available and, while complex to implement, would provide several advantages over our current voting system. It would eliminate the need for convenient polling locations, as you can vote anywhere. The timing of elections would be fine, as you could vote during a work break. The frequency of elections would also be less of an issue because you wouldn't have to go to a polling location. The complexity of the ballot could be addressed in the application, which would allow links showing descriptions of candidates and issues.
While voter turnout in the United States is low compared to other democratic countries, there are both simple and complex solutions that can increase voter participation. States with high voter turnout should be models for other states to follow as they have implemented practical reforms like same-day voter registration. The United States can then look to other countries and enact common sense solutions like voting on the weekend or on election holidays. Finally, we can use technology safely and responsibly to take voting into the 21st century and remove all the barriers to voting that our existing system places between the voter and the ballot.
|
|
null | false
| null |
categorize some popular sports as either 'safe', 'not so safe' or 'dangerous'.
|
safe: pingpong, golf, chess, curling, cross county skiing, archery
not so safe: soccer, basketball, bicycling, baseball, tenniss
dangerous: skiing, surfing, American Football, race car driving, sky diving
|
|
null | false
| null |
Which female artists sold over 40 million copies of an album?
|
Only 2 female artists sold over 40 million copies of an album.
- Whitney Houston / various artists "The Bodyguard" in 1992
- Shania Twain "Come On Over" in 1997
|
|
null | false
| null |
Polly, a 51-year-old female, visits your fitness facility. Polly has been told she has pre-diabetes. Explain to her what this means. What role can exercise play in reversing this?
|
From a scientific point of view, this means that she is insulin resistant, but she does not have hyperglycemia. As a result, she should be concerned; she is at a high risk for developing type 2 diabetes. The health risks are numerous: damage to blood vessels, high blood pressure, obesity, osteoporosis (thinning bones), and even certain types of cancer like colon, breast, and prostate. However, she can lower insulin resistance and mitigate her risk of pre-diabetes by improving her exercise habits. According to ACSM’s Health and Fitness Journal, almost “all physical activity has a positive effect on insulin.” To maximize efficacy, she should focus on moderate to higher intensity physical activities to best enhance insulin action and glycemic control.
|
|
null | false
| null |
The nominees for the 70th Academy Awards were announced on February 10, 1998, at the Samuel Goldwyn Theater in Beverly Hills, California, by Robert Rehme, president of the academy, and actress Geena Davis. Titanic received the most nominations with a record-tying fourteen (1950's All About Eve, and later 2016's La La Land, also achieved this distinction); Good Will Hunting and L.A. Confidential came in second with nine apiece.
|
What movie received the most nominations for the 70th Academy Awards?
|
Titanic received the most nominations at fourteen nominations at the 70th Academy Awards
|
null | false
| null |
Who wrote the book "The 7 Habits of Highly Effective Families" ?
|
Stephen R. Covey
|
|
null | false
| null |
Mount Shakspere is a 12,174-foot-elevation (3,711 meter) summit located in Fresno County, California, United States.
Description
The mountain is set four miles west of the crest of the Sierra Nevada mountain range, in the Palisades area of Kings Canyon National Park. It is situated 4.5 miles (7.2 km) southwest of Middle Palisade, 3.5 miles (5.6 km) southeast of Giraud Peak, and one mile north-northwest of line parent Observation Peak. Mt. Shakspere ranks as the 384th-highest summit in California, and topographic relief is significant as the summit rises over 3,500 feet (1,100 meters) above Palisade Creek in approximately one mile (1.6 km). The first ascent of the summit was made July 20, 1930, by Francis P. Farquhar, Mary Lou Michaels, Doris Drust, Lorna Kilgariff, and Robert L. Lipman. This mountain's name has been officially adopted by the United States Board on Geographic Names.
|
How tall is Mt. Shakspere?
|
Mount Shakspere is 12,174 feet tall
|
1910.04006
| true
| null |
To systematically evaluate the importance of the clinical sentiment values extracted from the free text in EHRs, we first build a baseline model using the structured features, which are similar to prior studies on readmission risk prediction BIBREF6. We then compare two models incorporating the unstructured features. In the "Baseline+Domain Sentences" model, we consider whether adding the counts of sentences per EHR that involve each of the seven risk factor domains as identified by our topic extraction model improved the model performance. In the "Baseline+Clinical Sentiment" model, we evaluate whether adding clinical sentiment scores for each risk factor domain improved the model performance. We also experimented with combining both sets of features and found no additional improvement.
To systematically evaluate the importance of the clinical sentiment values extracted from the free text in EHRs, we first build a baseline model using the structured features, which are similar to prior studies on readmission risk prediction BIBREF6.
|
Do they compare to previous models?
|
Yes.
|
1910.00825
| false
| null |
In this case, the summary has a high ROUGE score, as it has a considerable proportion of word overlap with the reference summary. However, it still has poor relevance and readability, for leaving out one of the most critical information: [time]. ROUGE treats each word equally in computing n-gram overlap while the informativeness actually varies: common words or phrases (e.g. “You are going to") significantly contribute to the ROUGE score and readability, but they are almost irrelevant to essential contents. The semantic slot values (e.g. [restaurant_name], [time]) are more essential compared to other words in the summary. However, ROUGE did not take this into consideration. To address this drawback in ROUGE, we propose a new evaluation metric: Critical Information Completeness (CIC). Formally, CIC is a recall of semantic slot information between a candidate summary and a reference summary. CIC is defined as follows:
where $V$ stands for a set of delexicalized values in the reference summary, $Count_{match}(v)$ is the number of values co-occurring in the candidate summary and reference summary, and $m$ is the number of values in set $V$. In our experiments, CIC is computed as the arithmetic mean over all the dialog domains to retain the overall performance.
CIC is a suitable complementary metric to ROUGE because it accounts for the most important information within each dialog domain. CIC can be applied to any summarization task with predefined essential entities. For example, in news summarization the proper nouns are the critical information to retain.
To address this drawback in ROUGE, we propose a new evaluation metric: Critical Information Completeness (CIC). Formally, CIC is a recall of semantic slot information between a candidate summary and a reference summary. CIC is defined as follows:
where $V$ stands for a set of delexicalized values in the reference summary, $Count_{match}(v)$ is the number of values co-occurring in the candidate summary and reference summary, and $m$ is the number of values in set $V$. In our experiments, CIC is computed as the arithmetic mean over all the dialog domains to retain the overall performance.
CIC is a suitable complementary metric to ROUGE because it accounts for the most important information within each dialog domain. CIC can be applied to any summarization task with predefined essential entities.
|
How does new evaluation metric considers critical informative entities?
|
Answer with content missing: (formula for CIC) it accounts for the most important information within each dialog domain. CIC can be applied to any summarization task with predefined essential entities
|
null | false
| null |
The Indian subcontinent is a physiographical region in Southern Asia, situated on the Indian Plate, projecting southwards into the Indian Ocean from the Himalayas.
|
what is the indian subcontinent
|
The Indian subcontinent is a physiographical region in Southern Asia, situated on the Indian Plate, projecting southwards into the Indian Ocean from the Himalayas. Geopolitically, it includes the countries of Bangladesh, Bhutan, India, Maldives, Nepal, Pakistan, and Sri Lanka. The terms "Indian subcontinent" and "South Asia" are often used interchangeably to denote the region, although the geopolitical term of South Asia frequently includes Afghanistan, which may otherwise be classified as Central Asian.
|
null | false
| 33
|
We build an INLINEFORM0 similarity matrix using an external corpus where the rows and columns represent words within the corpus and the element contains the similarity score between the row word and column word using the similarity measures discussed above. If a word maps to more than one possible sense, we use the sense that returns the highest similarity score.
For this paper our external corpus was the NLM 2015 Medline baseline. Medline is a bibliographic database containing over 23 million citations to journal articles in the biomedical domain and is maintained by National Library of Medicine. The 2015 Medline Baseline encompasses approximately 5,600 journals starting from 1948 and contains 23,343,329 citations, of which 2,579,239 contain abstracts. In this work, we use Medline titles and abstracts from 1975 to present day. Prior to 1975, only 2% of the citations contained an abstract. We then calculate the similarity for each bigram in this dataset and include those that have a similarity score greater than a specified threshold on these experiments.
In this work, we use Medline titles and abstracts from 1975 to present day. Prior to 1975, only 2% of the citations contained an abstract.
|
Why does the team choose 1975 as the starting year when using Medline titles and abstracts from NLM 2015 Medline baseline?
|
Because before the year 1975, only 2% of the citations contained an abstract.
|
null | false
| null |
What does it mean for a function to be computable?
|
Although informally a function is computable if its values can be calculated, Alan Turing, Alonzo Church, and Kurt Godel each had formal definitions of computablility, namely Turing Machines, Lambda Calculus, and general recursive functions. When it was shown in the Church-Turing thesis that these three formulations are equivalent, it became more clear that they could be used as a formal definition of computability.
|
|
2001.09332
| true
| null |
To analyse the results we chose to use the test provided by BIBREF10, which consists of $19\,791$ analogies divided into 19 different categories: 6 related to the “semantic" macro-area (8915 analogies) and 13 to the “syntactic" one (10876 analogies). All the analogies are composed by two pairs of words that share a relation, schematized with the equation: $a:a^{*}=b:b^{*}$ (e.g. “man : woman = king : queen"); where $b^{*}$ is the word to be guessed (“queen"), $b$ is the word coupled to it (“king"), $a$ is the word for the components to be eliminated (“man"), and $a^{*}$ is the word for the components to be added (“woman").
To analyse the results we chose to use the test provided by BIBREF10, which consists of $19\,791$ analogies divided into 19 different categories: 6 related to the “semantic" macro-area (8915 analogies) and 13 to the “syntactic" one (10876 analogies). All the analogies are composed by two pairs of words that share a relation, schematized with the equation: $a:a^{*}=b:b^{*}$ (e.g. “man : woman = king : queen"); where $b^{*}$ is the word to be guessed (“queen"), $b$ is the word coupled to it (“king"), $a$ is the word for the components to be eliminated (“man"), and $a^{*}$ is the word for the components to be added (“woman").
|
Are the word embeddings tested on a NLP task?
|
Yes.
|
1909.05360
| false
| null |
The TB-Dense dataset mitigates this issue by forcing annotators to examine all pairs of events within the same or neighboring sentences, and it has been widely evaluated on this task BIBREF3, BIBREF4, BIBREF19, BIBREF5. Recent data construction efforts such as MATRES BIBREF25 further enhance the data quality by using a multi-axis annotation scheme and adopting a start-point of events to improve inter-annotator agreements. We use TB-Dense and MATRES in our experiments and briefly summarize the data statistics in Table TABREF33.
We use TB-Dense and MATRES in our experiments and briefly summarize the data statistics in Table TABREF33.
|
What datasets were used for this work?
|
The answers are shown as follows:
* TB-Dense
* MATRES
|
null | false
| null |
South Park centers around four boys: Stan Marsh, Kyle Broflovski, Eric Cartman and Kenny McCormick. The boys live in the fictional small town of South Park, located within the real-life South Park basin in the Rocky Mountains of central Colorado, approximately a one-hour drive from Denver. The town is also home to an assortment of other characters, including students, families, elementary school staff, and other various residents. Prominent settings include South Park Elementary, various neighborhoods and the surrounding mountain range, actual Colorado landmarks, and the businesses along the town's main street, all of which are based on the appearance of similar locations in Fairplay, Colorado. As one of the few television programs set in the Mountain West region that takes place outside the urban core of Denver, South Park frequently features the unique culture of the region, including cattle ranchers, Old West theme parks, snowy climates, mountaineering, Mormons, real-life Colorado locations such as Casa Bonita and Cave of the Winds, and many other regionally specific characteristics.
Stan is portrayed as the everyman of the group, described in official media as an "average, American 4th grader." Kyle is Jewish, and his portrayal as one of the few such people in South Park is often dealt with satirically. Stan is modeled after Parker, while Kyle is modeled after Stone. They are best friends, and their friendship, symbolically intended to reflect Parker and Stone's friendship, is a common topic throughout the series. Cartman (as he is commonly referred to) is amoral and increasingly psychopathic, and is commonly portrayed as an antagonist. His staunch antisemitism has resulted in a progressive rivalry with Kyle. Kenny, who comes from a poor family, tightly wears his parka hood to the point where it obscures most of his face and muffles his speech. During the first five seasons, Kenny died in almost every episode before reappearing in the next with no definite explanation. He was killed off in the fifth season episode "Kenny Dies", being reintroduced in the sixth season finale. Since then, Kenny is depicted as dying sporadically. During the first 58 episodes, the children were in the third grade. During the fourth season, they entered the fourth grade, where they have remained ever since.
Plots are often set in motion by events, ranging from the fairly typical to the supernatural and extraordinary, which frequently happen in the town. The boys often act as the voice of reason when these events cause panic or incongruous behavior among the adult populace, who are customarily depicted as irrational, gullible, and prone to overreaction. They are frequently confused by the contradictory and hypocritical behavior of their parents and other adults, and often perceive them as having distorted views on morality and society.
|
Without quoting directly from the text, give me a summary of what the show "South Park" is about.
|
"South Park" is a comedic cartoon about four boys who live in a fictitious town called South Park, Colorado. The four boys are best friends and go on various adventures, where they run into situations that tackle moral and hypocritical situations they see from the adults in town. The show is a satire and often will center around a prevelant worldly topic.
|
null | false
| null |
Identify which instrument is string or percussion: Tom-tom drum, Viola de arame
|
Viola de arame is string, Tom-tom drum is percussion.
|
|
null | false
| null |
Rajendra Tower (Hindi: राजेंद्र टावर), locally known as Tower Chowk, is a historic clock tower in Gaya, Bihar, India. Built sometime between 1910 and 1914 during the British rule by the-then collector George Oldhum, it was renamed as Rajendra Tower after the first Indian president Dr. Rajendra Prasad in 1981 by the administration of Bihar's Chief Minister Jagannath Mishra. The tower is widely considered to be the symbol of the city.
History
The clock tower was built during the British Raj by George Oldhum, the-then collector of Gaya district, which was under Bihar and Orissa Province sometime between 1910 and 1914. The clock tower gradually became symbolic for the city of Gaya. After the end of British rule in 1947, the clock tower eventually became the property of the new Indian state of Bihar. It was renamed Rajendra Tower in 1981 after Dr. Rajendra Prasad, the first President of India by the administration of Jagannath Mishra, the-then Chief Minister of Bihar. According to locals, there were four clocks in the tower on each side and people used to hear the chimes of the clock from meters away whereas pedestrians used to match their accurate time by taking the reference of the clock. Due to the presence of the clock tower, the area gradually became a commercial district and goes by the name of Tower Chowk.
|
What was the Tower Chowk renamed?
|
The tower Chowk was renamed as Rajendra Tower after the passing of the president Dr. Rajendra Prasad.
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.