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
| null |
Where do we find Asiatic lions in India?
|
Gujarat is the only state in India with a Lion population and Ghir national forest is home to many of the Asiatic lions. Lions were protected in India after Governments started establishing National reserve forests across the country.
|
|
null | false
| null |
Write a haiku about sitting on the shore and watching the waves.
|
I sit on the shore
observing the waves crashing
over and over
|
|
null | false
| null |
How do you calculate a golfer's handicap?
|
A golfer's handicap is calculated by averaging the best 8 scores out of the last 20 rounds.
|
|
null | false
| null |
Is the headquarters of Surrey County Council in Surrey?
|
In 2021 Surrey County Council in England finally moved its headquarters back into its area of jurisdiction from its former location in the Royal Borough of Kingston upon Thames. The new headquarters, in the Surrey town of Reigate, are on a site previously occupied by Canon UK. Historically Kingston had been part of the county of Surrey until it was made part of Greater London in 1965. Postal addresses in Kingston still often include Surrey despite not being part of the county anymore.
|
|
null | false
| null |
What is the name of the movie made by Ben Affleck and Matt Damon about the company Nike?
|
Air
|
|
null | false
| 138
|
Many studies have focused on analyzing Flickr tags to extract useful information in domains such as linguistics BIBREF40 , geography BIBREF0 , BIBREF41 , and ecology BIBREF42 , BIBREF7 , BIBREF43 . Most closely related to our work, BIBREF7 found that the tags of georeferenced Flickr photos can effectively supplement traditional scientific environmental data in tasks such as predicting climate features, land cover, species occurrence, and human assessments of scenicness. To encode locations, they simply combine a bag-of-words representation of geographically nearby tags with a feature vector that encodes associated structured scientific data. They found that the predictive value of Flickr tags is roughly on a par with that of the scientific datasets, and that combining both types of information leads to significantly better results than using either of them alone. As we show in this paper, however, their straightforward way of combining both information sources, by concatenating the two types of feature vectors, is far from optimal.
Despite the proven importance of Flickr tags, the problem of embedding Flickr tags has so far received very limited attention. To the best of our knowledge, BIBREF44 is the only work that generated embeddings for Flickr tags. However, their focus was on learning embeddings that capture word meaning (being evaluated on word similarity tasks), whereas we use such embeddings as part of our method for representing locations.
Many studies have focused on analyzing Flickr tags to extract useful information in domains such as linguistics [9], geography [5,14], and ecology [1,21,22]. Most closely related to our work, [21] found that the tags of georeferenced Flickr photos can effectively supplement traditional scientific environmental data in tasks such as predicting climate features, land cover, species occurrence, and human assessments of scenicness.
|
What advantages do the tags of georeferenced Flickr photos according to others' research?
|
It can effectively supplement traditional scientific environmental data in tasks such as predicting climate features, land cover, species occurrence, and human assessments of scenicness.
|
null | false
| null |
What is the population of Canada?
|
Canada is a country in North America with a population of 38.25 million people. Its ten provinces and three territories extend from the Atlantic Ocean to the Pacific Ocean and northward into the Arctic Ocean, making it the world's second-largest country by total area, with the world's longest coastline. It is characterized by a wide range of both meteorologic and geological regions. The country is sparsely inhabited, with the vast majority residing south of the 55th parallel in urban areas. Canada's capital is Ottawa and its three largest metropolitan areas are Toronto, Montreal, and Vancouver.
|
|
null | false
| null |
Artemis I was formerly known as
|
Exploration Mission-1
|
|
null | false
| 7
|
The Flickr30K dataset BIBREF0 is a collection of over 30,000 images with 5 crowdsourced descriptions each. It is commonly used to train and evaluate neural network models that generate image descriptions (e.g. BIBREF2 ). An untested assumption behind the dataset is that the descriptions are based on the images, and nothing else. Here are the authors (about the Flickr8K dataset, a subset of Flickr30K):
“By asking people to describe the people, objects, scenes and activities that are shown in a picture without giving them any further information about the context in which the picture was taken, we were able to obtain conceptual descriptions that focus only on the information that can be obtained from the image alone.” BIBREF1
What this assumption overlooks is the amount of interpretation or recontextualization carried out by the annotators. Let us take a concrete example. Figure FIGREF1 shows an image from the Flickr30K dataset.
This image comes with the five descriptions below. All but the first one contain information that cannot come from the image alone. Relevant parts are highlighted in bold:
We need to understand that the descriptions in the Flickr30K dataset are subjective descriptions of events. This can be a good thing: the descriptions tell us what are the salient parts of each image to the average human annotator. So the two humans in Figure FIGREF1 are relevant, but the two soap dispensers are not. But subjectivity can also result in stereotypical descriptions, in this case suggesting that the male is more likely to be the manager, and the female is more likely to be the subordinate. rashtchian2010collecting do note that some descriptions are speculative in nature, which they say hurts the accuracy and the consistency of the descriptions. But the problem is not with the lack of consistency here. Quite the contrary: the problem is that stereotypes may be pervasive enough for the data to be consistently biased. And so language models trained on this data may propagate harmful stereotypes, such as the idea that women are less suited for leadership positions.
This paper aims to give an overview of linguistic bias and unwarranted inferences resulting from stereotypes and prejudices. I will build on earlier work on linguistic bias in general BIBREF3 , providing examples from the Flickr30K data, and present a taxonomy of unwarranted inferences. Finally, I will discuss several methods to analyze the data in order to detect biases.
This paper aims to give an overview of linguistic bias and unwarranted inferences resulting from stereotypes and prejudices.
|
What does the paper aim to do?
|
To give an overview of linguistic bias and unwarranted inferences resulting from stereotypes and prejudices.
|
null | false
| 21
|
Many reinforcement learning algorithms are designed for relatively small discrete or continuous action spaces and so have trouble scaling. Text-adventure games—or interaction fictions—are simulations in which both an agents' state and action spaces are in textual natural language. An example of a one turn agent interaction in the popular text-game Zork1 can be seen in Fig. FIGREF1. Text-adventure games provide us with multiple challenges in the form of partial observability, commonsense reasoning, and a combinatorially-sized state-action space. Text-adventure games are structured as long puzzles or quests, interspersed with bottlenecks. The quests can usually be completed through multiple branching paths. However, games can also feature one or more bottlenecks. Bottlenecks are areas that an agent must pass through in order to progress to the next section of the game regardless of what path the agent has taken to complete that section of the quest BIBREF0. In this work, we focus on more effectively exploring this space and surpassing these bottlenecks—building on prior work that focuses on tackling the other problems.
Formally, we use the definition of text-adventure games as seen in BIBREF1 and BIBREF2. These games are partially observable Markov decision processes (POMDPs), represented as a 7-tuple of $\langle S,T,A,\Omega , O,R, \gamma \rangle $ representing the set of environment states, mostly deterministic conditional transition probabilities between states, the vocabulary or words used to compose text commands, observations returned by the game, observation conditional probabilities, reward function, and the discount factor respectively. For our purposes, understanding the exact state and action spaces we use in this work is critical and so we define each of these in relative depth.
Action-Space. To solve Zork1, the cannonical text-adventure games, requires the generation of actions consisting of up to five-words from a relatively modest vocabulary of 697 words recognized by the game’s parser. This results in $\mathcal {O}(697^5)={1.64e14}$ possible actions at every step. To facilitate text-adventure game playing, BIBREF2 introduce Jericho, a framework for interacting with text-games. They propose a template-based action space in which the agent first selects a template, consisting of an action verb and preposition, and then filling that in with relevant entities $($e.g. $[get]$ $ [from] $ $)$. Zork1 has 237 templates, each with up to two blanks, yielding a template-action space of size $\mathcal {O}(237 \times 697^2)={1.15e8}$. This space is still far larger than most used by previous approaches applying reinforcement learning to text-based games.
State-Representation. Prior work has shown that knowledge graphs are effective in terms of dealing with the challenges of partial observability $($BIBREF3 BIBREF3; BIBREF4$)$. A knowledge graph is a set of 3-tuples of the form $\langle subject, relation, object \rangle $. These triples are extracted from the observations using Stanford's Open Information Extraction (OpenIE) BIBREF5. Human-made text-adventure games often contain relatively complex semi-structured information that OpenIE is not designed to parse and so they add additional rules to ensure that the correct information is parsed. The graph itself is more or less a map of the world, with information about objects' affordances and attributes linked to the rooms that they are place in a map. The graph also makes a distinction with respect to items that are in the agent's possession or in their immediate surrounding environment. An example of what the knowledge graph looks like and specific implementation details can be found in Appendix SECREF14.
BIBREF6 introduce the KG-A2C, which uses a knowledge graph based state-representation to aid in the section of actions in a combinatorially-sized action-space—specifically they use the knowledge graph to constrain the kinds of entities that can be filled in the blanks in the template action-space. They test their approach on Zork1, showing the combination of the knowledge graph and template action selection resulted in improvements over existing methods. They note that their approach reaches a score of 40 which corresponds to a bottleneck in Zork1 where the player is eaten by a “grue” (resulting in negative reward) if the player has not first lit a lamp. The lamp must be lit many steps after first being encountered, in a different section of the game; this action is necessary to continue exploring but doesn’t immediately produce any positive reward. That is, there is a long term dependency between actions that is not immediately rewarded, as seen in Figure FIGREF1. Others using artificially constrained action spaces also report an inability to pass through this bottleneck BIBREF7, BIBREF8. They pose a significant challenge for these methods because the agent does not see the correct action sequence to pass the bottleneck enough times. This is in part due to the fact that for that sequence to be reinforced, the agent needs to reach the next possible reward beyond the bottleneck.
More efficient exploration strategies are required to pass bottlenecks. Our contributions are two-fold. We first introduce a method that detects bottlenecks in text-games using the overall reward gained and the knowledge graph state. This method freezes the policy used to reach the bottleneck and restarts the training from there on out, additionally conducting a backtracking search to ensure that a sub-optimal policy has not been frozen. The second contribution explore how to leverage knowledge graphs to improve existing exploration algorithms for dealing with combinatorial action-spaces such as Go-Explore BIBREF9. We additionally present a comparative ablation study analyzing the performance of these methods on the popular text-game Zork1.
They test their approach on Zork1, showing the combination of the knowledge graph and template action selection resulted in improvements over existing methods.
|
What result does Ammanabrolu & Hausknecht get after they test their approach on Zork1?
|
The combination of the knowledge graph and template action selection resulted in improvements over existing methods.
|
null | false
| 191
|
Traditional approaches BIBREF0 , BIBREF1 , BIBREF2 for sentence relation modeling tasks such as paraphrase identification, question answering, recognized textual entailment and semantic textual similarity prediction usually build the supervised model using a variety of hand crafted features. Hundreds of features generated at different linguistic levels are exploited to boost classification. With the success of deep learning, there has been much interest in applying deep neural network based techniques to further improve the prediction performances BIBREF3 , BIBREF4 , BIBREF5 .
A key component of deep neural network is word embedding which serve as an lookup table to get word representations. From low level NLP tasks such as language modeling, POS tagging, name entity recognition, and semantic role labeling BIBREF6 , BIBREF7 , to high level tasks such as machine translation, information retrieval and semantic analysis BIBREF8 , BIBREF9 , BIBREF10 . Deep word representation learning has demonstrated its importance for these tasks. All the tasks get performance improvement via further learning either word level representations or sentence level representations. On the other hand, some researchers have found character-level convolutional networks BIBREF11 , BIBREF12 are useful in extracting information from raw signals for the task such as language modeling or text classification.
In this work, we focus on deep neural network based sentence relation modeling tasks. We explore treating each sentence as a kind of raw signal at character level, and applying temporal (one-dimensional) Convolution Neural Network (CNN) BIBREF6 , Highway Multilayer Perceptron (HMLP) and multi-layer bidirectional LSTM (Long Short Term Memory) BIBREF13 to learn sentence representations. We propose a new deep neural network architecture that jointly leverage pre-trained word embedding and character embedding to represent the meaning sentences. More specifically, our new approach first generates two kinds of word sequence representations. One kind of sequence representations are the composition of pre-trained word vectors. The other kind of sequence representation comprise word vectors that generating from character-level convolutional network. We then inject the two sequence representations into bidirectional LSTM, which means forward directional LSTM accept pre-trained word embedding output and backward directional LSTM accept auxiliary character CNN embedding output. The final sentence representation is the concatenation of the two direction. After that, we construct matching features followed by another temporal CNN to learn high-level hidden matching feature representations. Figure FIGREF1 shows the neural network architecture for general sentence relation modeling.
Our model shows that when trained on small size datasets, combining pre-trained word embeddings with auxiliary character-level embedding can improve the sentence representation. Word embeddings can help capturing general word semantic meanings, whereas char-level embedding can help modeling task specific word meanings. Note that auxiliary character-level embedding based sentence representation do not require the knowledge of words or even syntactic structure of a language. The enhanced sentence representation generated by multi-layer bidirectional LSTM will encapsulate the character and word levels informations. Furthermore, it may enhance matching features that generated by computing similarity measures on sentence pairs. Quantitative evaluations on standard dataset demonstrate the effectiveness and advantages of our method.
Our model shows that when trained on small size datasets, combining pre-trained word embeddings with auxiliary character-level embedding can improve the sentence representation.
|
According to authors' findings, when trained on small size datasets, what can be done to improve the sentence representation?
|
Combining pre-trained word embeddings with auxiliary character CNN embedding output.
|
null | false
| null |
What is the difference between archaeology and paleontology?
|
Archaeology is the study of human culture, while paleontology is the study of dinosaurs.
|
|
null | false
| null |
Tell me where the following numbers are odd or even: 13, 200, 334, 420, 5, 12
|
13: Odd
200: Even
334: Even
420: Even
5: Odd
12: Even
|
|
null | false
| null |
Give me a list of 10 commonly used seasonings in Chinese cooking
|
Soy Sauce
Oyster Sauce
Chinese Cooking Wine
Garlic
Green Onion
Ginger
Dark Soy Sauce
Sichuan peppercorn
White Pepper
Sesame Oil
|
|
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.
The transformation of a set of documents into a short text containing their most important aspects, is a long-studied problem in NLP.
|
What's the long-studied problem in NLP?
|
The transformation of a set of documents into a short text containing their most important aspects.
|
null | false
| null |
Which of the following would be considered assets on a financial statement: Inventory, Revenue, Leases, Accounts Payable, Retained Earnings, Cash, Treasury Bills
|
Inventory, Cash, Treasury Bills
|
|
null | false
| 154
|
All corpora provide datasets/splits for answer selection, whereas only (WikiQA, SQuAD) and (WikiQA, SelQA) provide datasets for answer extraction and answer triggering, respectively. SQuAD is much larger in size although questions in this corpus are often paraphrased multiple times. On the contrary, SQuAD's average candidates per question ( INLINEFORM0 ) is the smallest because SQuAD extracts answer candidates from paragraphs whereas the others extract them from sections or infoboxes that consist of bigger contexts. Although InfoboxQA is larger than WikiQA or SelQA, the number of token types ( INLINEFORM1 ) in InfoboxQA is smaller than those two, due to the repetitive nature of infoboxes.
All corpora show similar average answer candidate lengths ( INLINEFORM0 ), except for InfoboxQA where each line in the infobox is considered a candidate. SelQA and SQuAD show similar average question lengths ( INLINEFORM1 ) because of the similarity between their annotation schemes. It is not surprising that WikiQA's average question length is the smallest, considering their questions are taken from search queries. InfoboxQA's average question length is relatively small, due to the restricted information that can be asked from the infoboxes. InfoboxQA and WikiQA show the least question-answer word overlaps over questions and answers ( INLINEFORM2 and INLINEFORM3 in Table TABREF2 ), respectively. In terms of the F1-score for overlapping words ( INLINEFORM4 ), SQuAD gives the least portion of overlaps between question-answer pairs although WikiQA comes very close.
Fig. FIGREF4 shows the distributions of seven question types grouped deterministically from the lexicons. Although these corpora have been independently developed, a general trend is found, where the what question type dominates, followed by how and who, followed by when and where, and so on.
Fig. FIGREF6 shows the distributions of answer categories automatically classified by our Convolutional Neural Network model trained on the data distributed by li:02a. Interestingly, each corpus focuses on different categories, Numeric for WikiQA and SelQA, Entity for SQuAD, and Person for InfoboxQA, which gives enough diversities for statistical learning to build robust models.
SQUAD provides no mapping of the answer contexts to Wikipedia, whereas WIKIQA and SELQA provide mappings; however, their data do not come from the same version of Wikipedia.
|
Does the SQUAD provide the mapping of the answer contexts to Wikipedia?
|
No, it doesn't.
|
null | false
| 451
|
A kernel machine uses the kernel trick to map the non-linear problem into a feature space where the problem may be linearly separable with an appropriate kernel function. Next we give the formal definition of kernel machines discussed in this paper. Given a training data set {X, y} of n training instances where {X ∈ R n×d , y ∈ R n } = {(x 1 , y 1 ), (x 2 , y 2 ), ..., (x n , y n )}, and (x i , y i ) denotes the instance x i ∈ R d with its label y i , the objective of the kernel machine training is to find an optimal ω * which minimizes the structural risk as follows.
min
where λ denotes the regularization constant and f (ω, x i ) = ω, φ(x i ) . The variable ω is defined on the reproducing kernel Hilbert space (RKHS) and •, • is the inner product on the RKHS. The function φ(•) maps the instances from their original data space to a higher dimensional feature space induced by the kernel function. Assume the loss l(•, •) is an affine function of ω. The representer theorem shows that a minimizer of the optimization problem ()
where k(x i , x j ) denotes a positive definite kernel function and k(x i , x j ) = φ(x i ), φ(x j ) . By substituting the expressions of f (ω, x i ) and ω into the Equation (1), we have the objective with respect to α below.
where α = [α 1 . . . α n ] T is an n-dimension vector, each dimension of which corresponds to the contribution of a training instance to the kernel machine.
Then, we can derive that the Hessian matrix H = [H ij ] n×n of Problem () is equal to the kernel matrix. The element in the i-th row and j-th column of the matrix H is H ij = k(x i , x j ). The derivations of the first and second derivatives can be found in the supplementary material. For clarity, we use kernel matrix H to denote the Hessian matrix of the kernel machine in the rest of the paper.
For clarity, we use kernel matrix H to denote the Hessian matrix of the kernel machine in the rest of the paper.
|
Is the hessian matrix mentioned in the discussion of SGD referring to kernel matrix?
|
We have replaced the Hessian matrix with kernel matrix.
|
1610.09516
| true
| null |
Street gang members have established online presences coinciding with their physical occupation of neighborhoods. The National Gang Threat Assessment Report confirms that at least tens of thousands of gang members are using social networking websites such as Twitter and video sharing websites such as YouTube in their daily life BIBREF0 . They are very active online; the 2007 National Assessment Center's survey of gang members found that 25% of individuals in gangs use the Internet for at least 4 hours a week BIBREF4 . Gang members typically use social networking sites and social media to develop online respect for their street gang BIBREF5 and to post intimidating, threatening images or videos BIBREF6 . This “Cyber-” or “Internet banging” BIBREF7 behavior is precipitated by the fact that an increasing number of young members of the society are joining gangs BIBREF8 , and these young members have become enamored with technology and with the notion of sharing information quickly and publicly through social media. Stronger police surveillance in the physical spaces where gangs congregate further encourages gang members to seek out virtual spaces such as social media to express their affiliation, to sell drugs, and to celebrate their illegal activities BIBREF9 .
The National Gang Threat Assessment Report confirms that at least tens of thousands of gang members are using social networking websites such as Twitter and video sharing websites such as YouTube in their daily life BIBREF0 . They are very active online; the 2007 National Assessment Center's survey of gang members found that 25% of individuals in gangs use the Internet for at least 4 hours a week BIBREF4 .
|
Do the authors provide evidence that 'most' street gang members use Twitter to intimidate others?
|
No.
|
2002.09758
| false
| null |
Table shows how unsupervised decompositions affect QA. Our RoBERTa baseline performs quite well on HotpotQA (77.0 F1), despite processing each paragraph separately, which prohibits inter-paragraph reasoning. The result is in line with prior work which found that a version of our baseline QA model using BERT BIBREF26 does well on HotpotQA by exploiting single-hop reasoning shortcuts BIBREF21. We achieve significant gains over our strong baseline by leveraging decompositions from our best decomposition model, trained with USeq2Seq on FastText pseudo-decompositions; we find a 3.1 F1 gain on the original dev set, 11 F1 gain on the multi-hop dev set, and 10 F1 gain on the out-of-domain dev set. Unsupervised decompositions even match the performance of using (within our pipeline) supervised and heuristic decompositions from DecompRC (i.e., 80.1 vs. 79.8 F1 on the original dev set).
We achieve significant gains over our strong baseline by leveraging decompositions from our best decomposition model, trained with USeq2Seq on FastText pseudo-decompositions; we find a 3.1 F1 gain on the original dev set, 11 F1 gain on the multi-hop dev set, and 10 F1 gain on the out-of-domain dev set.
|
How large is the improvement over the baseline?
|
The answers are shown as follows:
* 3.1 F1 gain on the original dev set
* 11 F1 gain on the multi-hop dev set
* 10 F1 gain on the out-of-domain dev set.
|
null | false
| null |
What is an excavator?
|
An excavator is a heavy equipment machine that is used for moving large amount of earth at one time. They have a boom and a huge bucket attached to them and are essentially a huge shovel used to prepare the ground for huge construction projects. John Deere is one company that produces excavators.
|
|
null | false
| null |
Hallmarking dates back to the 14th century when Edward I of England passed a law requiring any item made of silver, which was offered for sale to be at least of equal quality as that of the coin of the realm (silver currency). The four wardens of the Goldsmiths’ Company were tasked with visiting workshops in the City of London to assay (test) silver articles. If these articles were found to be below standard they were originally forfeit to the king, but if they passed, each article received the king's mark of authentication which was the mark of a leopard's head. By 1478, there were several hundred workshops and merchants manufacturing silver articles in the City of London. It was not possible for the wardens to visit them all so the merchants were ordered to bring their items to Goldsmiths’ Hall for testing and marking and a permanent assay office was established in the building. This is the origin of the term hallmark – struck with the king's mark at Goldsmiths’ Hall.
In 1544 the Goldsmith's Company adopted the king's mark as their town mark and the mark of the leopard's head is now internationally recognized as the mark of this assay office.
|
From the passage provided, extract when hallmarking was first observed in England.
|
Hallmarking was first observed in England in the 14th century.
|
null | false
| null |
What is email?
|
Email is form of electronic communication that is distributed between computers over closed and open networks. The internet is the most common network that is used for email communication. Email services that are very popular are gmail, outook, protonmail, and yahoo mail. Major fortune 500 companies all use email communication in one form or another to help their companies and people all be more productive and deliver results faster.
|
|
null | false
| null |
Jimmy Cozier (born February 6, 1977) is an American R&B singer and songwriter. He is best known for his hit single "She's All I Got" and for being one of the inaugural artists signed to Clive Davis' J Records label.
Biography
The son of Guyanese American jazz saxophonist Jimmy Cozier, and Dawn Cozier, a Jamaican born hair stylist. Cozier and his younger brother Malik were raised in Crown Heights, Brooklyn. Cozier was encouraged to sing as a child by his family, who would demand that he perform in group settings. He started out as a singer/songwriter for artists such as Mýa, Sinéad O'Connor, and Janet Jackson (whose hit "Girlfriend/Boyfriend" he co-wrote). He was a background vocalist for the Junior Mafia/Lil' Kim track "Backstabbers" and toured with Joe behind the latter's album All That I Am.
Wyclef Jean caught word of Cozier's talent thru Cozier manager Jacques “Haitian Jack” Agnant and had him meet with Clive Davis, who signed Cozier to J Records in 2000. His debut single "She's All I Got" was released in 2001, and rose to #26 on the Billboard Hot 100 and to #4 on the R&B chart.
Following the success of the single, his self-titled debut album was released on July 9, 2001 and hit the Billboard Top 200 at #65 and #15 on the R&B Albums chart. A follow-up single "So Much to Lose" was released later in the year, and peaked at #123 on the R&B chart. He also appeared on labelmate Alicia Keys' debut Songs in A Minor on the track "Mr. Man" which he co-wrote, and "Turn Around" with other labelmate Olivia, on her self-titled debut album.
He began writing songs for other artists such as Chris Brown, Olivia and Cheri Dennis in the years since. After the folding of J Records, he became an independent artist, creating his own label, Cozi Music Inc.
In 2007, he announced work on a new album, releasing two singles, "U Got Them Goods" and "You", that year. Continued work on his sophomore album stretched through 2010.
His second album, Way of Life was announced in 2010, led by the single "Tonight" which was released on March 2, 2010. The album was released on March 16, 2010.
In 2013, he collaborated with Sean Paul for his single "Always Be My Lady" which was released on August 2. Additionally, he released a single "Girls Girls" that year.
The next year, he premiered a new single "Choose Me" featuring Shaggy and announced his third album would be titled R&B (Reggae & Blues) set for a 2015 release. Following the single, he collaborated with artists such as Gyptian, Elephant Man, Shaggy, Red Fox, Melissa Musique on various compilation albums. One of those collaborations, "Special Girl (Remix)" was released as a single in 2014.
His third album, titled Get Cozi was released on December 9, 2016. The set's second single "I'll Change" was announced that same month.
Discography
Albums
Jimmy Cozier (2001)
Way of Life (2010)
Get Cozi (2016)
Singles
2001: "She's All I Got"
2001: "So Much To Lose"
2007: "U Got Them Goods"
2007: "You"
2010: "Tonight"
2013: "Always Be My Lady" (with Sean Paul)
2013: "Girls Girls"
2014: "Special Girl (Remix)" (with Supa Hype, Gyptian)
2014: "Choose Me" (featuring Shaggy)
2016: "I'll Change"
2018: "Love Her"
|
List all the singles Jimmy Cozier has.
|
2001: "She's All I Got"
2001: "So Much To Lose"
2007: "U Got Them Goods"
2007: "You"
2010: "Tonight"
2013: "Always Be My Lady" (with Sean Paul)
2013: "Girls Girls"
2014: "Special Girl (Remix)" (with Supa Hype, Gyptian)
2014: "Choose Me" (featuring Shaggy)
2016: "I'll Change"
2018: "Love Her"
|
null | false
| null |
The Garshasp-nama is an epic poem by Asadi Tusi (died 1072/73). It has been described as one of the best epic poems in Persian literature, comparable to Shahnameh, by Ferdowsi, and the most important work of Asadi Tusi. Asadi Tusi completed the poem in 1066 and dedicated it to a certain Abu Dulaf, the ruler of Nakhjavan (nothing is known about him). The poem has also been translated to other languages such as French and German. It seems that Asadi Tusi wrote this poem based on a written source. Like the Shahnameh, it contains few Arabic loan-words and consists of some 9,000 verses. The main hero of this epic poem is Garshasp, the son of Etret, and grandson of Sām. The poem begins with the story of Jamshid and Zahhak. Jamshid is overthrown by Zahhak and flees to Zabolistan. In Zabolistan, Jamshid falls in love with an unnamed daughter of Kurang, the king of Zabolistan, and she bore a child for Jamshid, named Tur (not to be confused with Tur, the son of Freydun). Jamshid flees again to China. Garshasp is actually the grandson of Tur's grandson.
|
Write down some key points for the poem Garshasp-nama based on the given text.
|
1. The Garshasp-nama is an epic poem by Asadi Tusi (died 1072/73).
2. It has been described as one of the best epic poems in Persian literature, comparable to Shahnameh, by Ferdowsi, and the most important work of Asadi Tusi.
3. Asadi Tusi completed the poem in 1066 and dedicated it to a certain Abu Dulaf, the ruler of Nakhjavan (nothing is known about him).
4. The poem has also been translated to other languages such as French and German.
5. It seems that Asadi Tusi wrote this poem based on a written source.
6. Like the Shahnameh, it contains few Arabic loan-words and consists of some 9,000 verses.
7. The main hero of this epic poem is Garshasp, the son of Etret, and grandson of Sām.
8. The poem begins with the story of Jamshid and Zahhak.
9. Jamshid is overthrown by Zahhak and flees to Zabolistan. In Zabolistan, Jamshid falls in love with an unnamed daughter of Kurang, the king of Zabolistan, and she bore a child for Jamshid, named Tur (not to be confused with Tur, the son of Freydun).
10. Jamshid flees again to China.
11. Garshasp is actually the grandson of Tur's grandson.
|
null | false
| null |
What are the pros and cons of working from home rather than going into an office downtown?
|
There are both advantages and disadvantages from working from home, and they apply differently to different people depending on the person's personality, family situation, location, and job description. Some key advantages are the ability to develop closer personal relationships with team members, have the opportunity for more spontaneous casual conversations, and experience more productive meetings because of the ability to read people's body language and to see presentation materials more clearly. Many people will experience better mental health by going into the office and establishing a sense of greater connection with team members. In addition, certain employees will take advantage of remote work to use work time to engage in personal activities rather than real work, to the detriment of the employer. Some disadvantages of going into the office include the time inefficiency from commuting, the cost of commuting, the time inefficiency that occurs from experiencing all the social chatting that occurs in the office, and the hassle of having to schedule conference rooms. Commuting also contributes to global warming through the greater use of fossil fuels. On balance, whether remote work makes sense is a complex question, and the answer will be unique for every employee and employer depending on the circumstances.
|
|
null | false
| null |
Name some famous poets and their poems.
|
Edgar Allen Poe
Birthplace: Boston
Famous poem: ”The Raven”
Famous quote: ”I have great faith in fools — self-confidence my friends call it.”
William Shakespeare
Birthplace: Stratford-upon-Avon, England
Famous poem: ”Sonnet XVIII” (Shall I compare thee to a summer’s day?)
Famous quote: ”All the world’s a stage, and all the men and women merely players: they have their exits and their entrances; and one man in his time plays many parts. His acts being seven ages.”
Maya Angelou
Birthplace: St. Louis
Famous poem: ”On the Pulse of Morning”
Famous quote: ”I’ve learned that people will forget what you said, people will forget what you did, but people will never forget how you made them feel.”
Emily Dickinson
Birthplace: Amherst, Massachusetts
Famous poem: “‘Hope’ is the thing with feathers”
Famous quote: “Hope is the thing with feathers that perches in the soul — and sings the tunes without the words — and never stops at all.”
Shel Silverstein
Birthplace: Chicago
Famous poem: “Where the Sidewalk Ends”
Famous quote: ”What I do is good. I wouldn’t let it out if I didn’t think it was.”
Robert Frost
Birthplace: San Francisco
Famous poem: “The Road Not Taken”
Famous quote: “The woods are lovely, dark and deep. But I have promises to keep, and miles to go before I sleep.”
Pablo Neruda
Birthplace: Parral, Chile
Famous poem: “I Do Not Love You Except Because I Love You”
Famous quote: “To feel the love of people whom we love is a fire that feeds our life.”
E. E. Cummings
Birthplace: Cambridge, Massachusetts
Famous poem: “i carry your heart with me”
Famous quote: “It takes courage to grow up and become who you really are.”
Langston Hughes
Birthplace: Joplin, Missouri
Famous poem: “I Too Sing America”
Famous quote: “Hold fast to dreams for if dreams die, life is a broken winged bird that cannot fly.”
Walt Whitman
Birthplace: Long Island, New York
Famous poem: “I Hear America Singing”
Famous quote: “Either define the moment or the moment will define you.”
Thomas Hardy
Birthplace: Dorset, England
Famous poem: “Hap”
Famous quote: “The main object of religion is not to get a man into heaven, but to get heaven into him.”
Rudyard Kipling
Birthplace: Bombay Presidency, British India
Famous poem: “Gunga Din”
Famous quote: “We have forty million reasons for failure, but not a single excuse.”
Oscar Wilde
Birthplace: Dublin, Ireland
Famous poem: “A Vision”
Famous quote: “I have the simplest tastes. I am always satisfied with the best.”
John Keats
Birthplace: London
Famous poem: “A Thing of Beauty (Endymion)”
Famous quote: “A thing of beauty is a joy forever; its loveliness increases; it will never pass into nothingness.”
Elizabeth Barrett Browning
Birthplace: Durham, England
Famous poem: “How Do I Love Thee?”
Famous quote: “If you desire faith, then you have faith enough.”
William Blake
Birthplace: London
Famous poem: “The Tyger”
Famous quote: “The glory of Christianity is to conquer by forgiveness.”
Sylvia Plath
Birthplace: Boston
Famous poem: “Daddy”
Famous quote: “Everything in life is writable about if you have the outgoing guts to do it and the imagination to improvise.”
Henry Wadsworth Longfellow
Birthplace: Portland, Maine
Famous poem: “The Song of Hiawatha”
Famous quote: “Give what you have. To someone, it may be better than you dare to think.”
William Wordsworth
Birthplace: Cumberland, England
Famous poem: “The Prelude”
Famous quote: “Fill your paper with the breathings of your heart.”
Mark Twain
Birthplace: Florida, Missouri
Famous poem: “Ode to Stephen Dowling Bots, Dec’d.”
Famous quote: “Kindness is the language which the deaf can hear and the blind can see.”
Ralph Waldo Emerson
Birthplace: Boston
Famous poem: “Uriel”
Famous quote: “A hero is no braver than an ordinary man, but he is braver five minutes longer.”
John Donne
Birthplace: London
Famous poem: “No Man Is An Island”
Famous quote: “Any man’s death diminishes me, because I am involved in Mankind; And therefore never send to know for whom the bell tolls; it tolls for thee.”
W.B. Yeats
Birthplace: County Dublin, Ireland
Famous poem: “The Second Coming”
Famous quote: “Education is not the filling of a pail, but the lighting of a fire.”
Lord Byron
Birthplace: London
Famous poem: “She Walks in Beauty”
Famous quote: “There is no instinct like that of the heart.”
Lewis Carroll
Birthplace: Cheshire, England
Famous poem: “Jabberwocky”
Famous quote: “It is one of the great secrets of life that those things which are most worth doing, we do for others.”
Alfred, Lord Tennyson
Birthplace: Lincolnshire, England
Famous poem: “The Charge of the Light Brigade”
Famous quote: “‘Tis better to have loved and lost Than never to have loved at all.”
Dante Alighieri
Birthplace: Florence, Italy
Famous poem: “Divine Comedy”
Famous quote: “Consider your origin; you were not born to live like brutes, but to follow virtue and knowledge.”
T.S. Eliot
Birthplace: St. Louis
Famous poem: “The Waste Land”
Famous quote: “Friendship should be more than biting time can sever.”
Ezra Pound
Birthplace: Hailey, Idaho
Famous poem: “Hugh Selwyn Mauberley”
Famous quote: “With one day’s reading a man may have the key in his hands.”
John Milton
Birthplace: London
Famous poem: “Paradise Lost”
Famous quote: “A good book is the precious lifeblood of a master spirit.”
Sappho
Birthplace: Lesbos, Greece
Famous poem: “Sappho 31”
Famous quote: “What is beautiful is good, and who is good will soon be beautiful.”
Homer
Birthplace: Smyrna (present-day Greece)
Famous poem: “The Iliad”
Famous quote: “Evil deeds do not prosper; the slow man catches up with the swift.”
Li Bai
Birthplace: Tang Empire (present-day Kyrgyzstan)
Famous poem: “Quiet Night Thought”
Famous quote: “We sit together, the mountain and me, until only the mountain remains.”
Jalal al-Din Rumi
Birthplace: Khorasan (present-day Afghanistan)
Famous poem: “Masnavi-ye Ma’navi (Spiritual Verses)”
Famous quote: “Raise your words, not voice. It is rain that grows flowers, not thunder.”
|
|
null | false
| 92
|
For the second round, we used a more powerful RoBERTa model BIBREF25 trained on SNLI, MNLI, an NLI-version of FEVER BIBREF26, and the training data from the previous round (A1). After a hyperparameter search, we selected the model with the best performance on the A1 development set. Then, using the hyperparameters selected from this search, we created a final set of models by training several models with different random seeds. During annotation, we constructed an ensemble by randomly picking a model from the model set as the adversary each turn. This helps us avoid annotators exploiting vulnerabilities in one single model. A new non-overlapping set of contexts was again constructed from Wikipedia via HotpotQA using the same method as Round 1.
For the second round, we used a more powerful RoBERTa model (Liu et al., 2019b) trained on SNLI, MNLI, an NLI-version2 of FEVER (Thorne et al., 2018), and the training data from the previous round (A1).
|
What did the authors use to train on SNLI, MNLI, and an NLI-version2 of FEVER in round 2?
|
A more powerful RoBERTa model.
|
null | false
| null |
Run Towards the Danger is a 2022 Canadian essay collection by Sarah Polley, a former child star, director, and screenwriter.
The six essays in the collection examine aspects of Polley's career on stage, screen, and on film detailing her roles in a Stratford Festival production of Alice Through the Looking Glass, as well as her breakout roles in The Adventures of Baron Munchausen and the TV series Road to Avonlea. The book also revealed for the first time that Polley had been a victim of Jian Ghomeshi who sexually and physically assaulted her when she was 16 and he was 28.
|
Given a reference text about Run Towards the Danger, tell me how many essays are part of the collection.
|
Six essays are part of the Run Towards the Danger essay collection.
|
1910.11204
| false
| null |
FLOAT SELECTED: Table 7: SRL results on the Chinese test set. We choose the best settings for each configuration of our model.
FLOAT SELECTED: Table 7: SRL results on the Chinese test set. We choose the best settings for each configuration of our model.
|
What are two strong baseline methods authors refer to?
|
Marcheggiani and Titov (2017) and Cai et al. (2018)
|
null | false
| null |
Can you summarize the book Annihilation?
|
Annihilation is a science fiction book taking place in the southern US inside a mysterious Area X. A team is assembled to explore the domain of Area X, led by a character only known as the psychologist. The main character is known as the biologist. These expeditions are managed by a mysterious government organization known as the Southern Reach.
The expedition centers on two notable landmarks within Area X, the topological anomaly, which the biologists calls a tower but is really a tunnel, and a lighthouse. The expedition encounters an entity within the topological anomaly that "writes" living words of non-sense on the wall of the anomaly in fungus.
Early in the expedition, the biologist is contaminated by the fungus of the words in the anomaly, which grants her immunity to hypnotic suggestion, which the psychologist uses to control the other members of the expedition. The plot follows the biologist pursuing the psychologist in search of answers, leading a confrontation at the lighthouse. The psychologist jumps from the lighthouse, presumably forced to by the entity behind Area X, but lives long enough to confront the biologist. During the climax of that encounter, the psychologist shouts "Annihilation!" multiple times, and it is revealed that this is a hypnotic trigger word designed to induce suicide in the other expedition members.
After the psychologist succumbs to her injuries at the base of the lighthouse, the biologist embarks on a journey north to an island that is possibly related to the disappearance of her husband, who was a member of the previous expedition.
|
|
null | false
| 44
|
We aimed to mimic the presence of sparse/noisy content distribution, mandating the need to curate a novel dataset via specific lexicons. We scraped 500 random posts from recognized transport forum. A pool of 50 uni/bi-grams was created based on tf-idf representations, extracted from the posts, which was further pruned by annotators. Querying posts on Twitter with extracted lexicons led to a collection of $19,300$ tweets. In order to have lexical diversity, we added 2500 randomly sampled tweets to our dataset. In spite of the sparse nature of these posts, the lexical characteristics act as information cues.
Figure FIGREF4 pictorially represents our methodology. Our approach required an initial set of informative tweets for which we employed two human annotators annotating a random sub-sample of the original dataset. From the 1500 samples, 326 were marked as informative and 1174 as non informative ($\kappa =0.81$), discriminated on this criteria: Is the tweet addressing any complaint or raising grievances about modes of transport or services/ events associated with transportation such as traffic; public or private transport?. An example tweet marked as informative: No, metro fares will be reduced ???, but proper fare structure needs to presented right, it's bad !!!.
We utilized tf-idf for the identification of initial seed phrases from the curated set of informative tweets. 50 terms having the highest tf-idf scores were passed through the complete dataset and based on sub-string match, the transport relevant tweets were identified. The redundant tweets were filtered based on the cosine similarity score. Implicit information indicators were identified based on domain relevance score, a metric used to gauge the coverage of n-gram (1,2,3) when evaluated against a randomly created pool of posts.
We collected a pool of 5000 randomly sampled tweets different from the data collection period. The rationale behind having such a metric was to discard commonly occurring n-grams normalized by random noise and include ones which are of lexical importance. We used terms associated with high domain relevance score (threshold determined experimentally) as seed phrases for the next set of iterations. The growing dictionary augments the collection process. The process ran for 4 iterations providing us 7200 transport relevant tweets as no new lexicons were identified. In order to identify linguistic signals associated with the complaint posts, we randomly sampled a set of 2000 tweets which was used as training set, manually annotated into distinct labels: complaint relevant (702) and complaint non-relevant (1298) ($\kappa =0.79$). We employed these features on our dataset.
Linguistic markers. To capture linguistic aspects of complaints, we utilized Bag of Words, count of POS tags and Word2vec clusters.
Sentiment markers. We used quantified score based on the ratio of tokens mentioned in the following lexicons: MPQA, NRC, VADER and Stanford.
Information specific markers. These account for a set of handcrafted features associated with complaint, we used the stated markers (a) Text-Meta Data, this includes the count of URL's, hashtags, user mentions, special symbols and user mentions, used to enhance retweet impact; (b) Request Identification, we employed the model presented in BIBREF3 to identify if a specific tweet assertion is a request; (c) Intensifiers, we make use of feature set derived from the number of words starting with capital letters and the repetition of special symbols (exclamation, questions marks) within the same post; (d) Politeness Markers, we utilize the politeness score of the tweet extracted from the model presented in BIBREF3; (e) Pronoun Variation, these have the ability to reveal the personal involvement or intensify involvement. We utilize the frequency of pronoun types $\lbrace \textit {first, second, third, demonstrative and indefinite}$} using pre-defined dictionaries.
From the pool of 7200 transport relevant tweets, we sampled 3500 tweets which were used as the testing set. The results are reported in TableTABREF5 with 10 fold cross-validation. With increasing the number of iterations, the pool of seed phrases gets refined and augments the selection of transport relevant tweets. The proposed pipeline is tailored to identify complaint relevant tweets in a noisy scenario.
The process ran for 4 iterations providing us 7200 transport relevant tweets as no new lexicons were identified.
|
How many times did the process run for iterations?
|
4.
|
null | false
| null |
Who was the mascot of 13th South Asian Games (SAG)
|
A pair of blackbucks
|
|
null | false
| 145
|
The goal of an language model is to assign meaningful probabilities to a sequence of words. Given a set of tokens $\mathbf {X}=(x_1,....,x_T)$, where $T$ is the length of a sequence, our task is to estimate the joint conditional probability $P(\mathbf {X})$ which is
were $(x_{1}, \ldots , x_{i-1})$ is the context. An Intrinsic evaluation of the performance of Language Models is perplexity (PPL) which is defined as the inverse probability of the set of the tokens and taking the $T^{th}$ root were $T$ is the number of tokens
In our two approaches we use transformer based architectures: BERT and Transformer-XL as mentioned before. Calculating the auto-regressive $P(\mathbf {X})$ for the transformer-XL is quite straight-forward as the model is unidirectional but it doesn't factorize the same way for a bi-directional model like BERT.
BERT's bi-directional context poses a problem for us to calculate an auto-regressive joint probability. A simple fix could be that we mask all the tokens $\mathbf {x}_{>i}$ and we calculate the conditional factors as we do for an unidirectional model. By doing so though, we loose upon the advantage of bi-directional context the BERT model enables. We propose an approximation of the joint probability as,
This type of approximations has been previously explored with Bi-directional RNN LM's BIBREF9 but not for deep transformer models. We therefore, define a pseudo-perplexity score from the above approximated joint probability.
The original BERT has two training objectives: 'Masked language modelling', in which you mask input tokens randomly and then predict the masked tokens using the left and right context. Additionally, there is the 'next sentence prediction' task that jointly trains text-pair representations. For training the Masked language model the original BERT used Byte Pair Encoding (BPE) BIBREF10 for subword tokenization BIBREF11.For example the rare word "unaffable" to be split up into more frequent subwords such as ["un", "##aff", "##able"]. To remain consistent with experiments performed with LSTM's we use the morfessor for the subword tokenization in the Finnish Language. In Addition, we also apply boundary markers as in (Table TABREF7) and train two separate models using this distinction. We train with left-marked markings as the original BERT was trained with such a scheme and the left+right-marked as it was the previous SOTA with the Finnish Language. For the transformer-XL experiments, we just train with the left+right marked scheme.
The Next Sentence Prediction (NSP) is a binary classification task which predicts whether two segments follow each other in the original text. This pre-training task was proposed to further improve the performance on downstreaming tasks, like Natural Language Inference(NLI) but in reality removing the NSP loss matches or slightly improves the downstream task performance BIBREF12. In this paper, we have omitted the NSP task from the BERT pre-training procedure and changed the input from a SEGMENT-PAIR input to a SINGLE SEGMENT input. As seen in (Fig FIGREF8)
Transformer-XL introduced the notion of recurrence in self-attention by caching the hidden state sequence to compute the hidden states of a new segment. It also introduces a novel relative positional embedding scheme and both of them combined address the issue of fixed context lengths. Transformer-XL as mentioned is a unidirectional deep transformer architecture, therefore the perplexity can be calculated as (Eq DISPLAY_FORM5). The only change is in the input format, were we use sub-word units rather than whole word units as Finnish is morphologically richer than English.
BERT’s bi-directional context poses a problem for us to calculate an auto-regressive joint probability.
|
What problem does BERT’s bi-directional context pose?
|
To calculate an auto-regressive joint probability.
|
null | false
| null |
The United Kingdom is a constitutional monarchy and parliamentary democracy. Its capital and largest city is London, the capital of England, a global city and financial centre with a population of over 14 million people. Edinburgh, Cardiff and Belfast are the national capitals of Scotland, Wales and Northern Ireland respectively.
|
Extract the names of the cities and rank them in alphabetical order?
|
The cities listed in alphabetical order are Belfast, Cardiff, Edinburgh and London.
|
1910.00825
| false
| null |
Moreover, we can easily extend SPNet to other summarization tasks. We plan to apply semantic slot scaffold to news summarization. Specifically, we can annotate the critical entities such as person names or location names to ensure that they are captured correctly in the generated summary. We also plan to collect a human-human dialog dataset with more diverse human-written summaries.
We plan to apply semantic slot scaffold to news summarization. Specifically, we can annotate the critical entities such as person names or location names to ensure that they are captured correctly in the generated summary. We also plan to collect a human-human dialog dataset with more diverse human-written summaries.
|
Is it expected to have speaker role, semantic slot and dialog domain annotations in real world datasets?
|
Not at the moment, but summaries can be additionaly extended with this annotations.
|
null | false
| null |
How is plastic made?
|
Plastic is made from oil, natural gas and even plant oils during refining of these oils into other products like gasoline. Ethane and propane are created when treated with heat during a refinery process called cracking. This turns the Ethane and propane into ethylene and propylene which are used with other chemical ingredients to create polymers that are the base of what plastic is made out of.
|
|
null | false
| null |
Which seventh season episode of Game of Thrones did Alan Taylor direct?
|
Alan Taylor directed "Beyond the Wall"
|
|
null | false
| null |
Where was the 2017 European Judo Championships held
|
The men's 90 kg competition at the 2017 European Judo Championships in Warsaw was held on 22 April at the Torwar Hall.
|
|
null | false
| null |
How can you open a can without using your hands?
|
1) You could try to use both feet to open the can if you have very strong feet.
2) You could ask a friend to open the can with their hands.
3) You could use your elbow to push the can off of a very high surface.
4) You could drive over the can with a car, though you would have to use only your arms to drive.
|
|
null | false
| null |
Ain Elzarga Wetland is a Ramsar-protected wetland of Libya. It was established in 2000 and covers an area of 33 hectares (82 acres). Ramsar says of it: "A small natural sebkha or depression with at least one natural connection to the sea, wet all year round but with increasing water levels and salinity during summer. The sebkha, with mudflats and salt marsh community, is surrounded by dunes from east to west and rocky hills to the south and east. The site is one of the most important wetlands in the area of the El Kouf National Park for migratory waterbirds. The birdwatching and ecotourism potential is considerable but undeveloped. Unsustainable hunting and destruction of vegetation, especially during summer, are considered threats.
|
Given this paragraph about the Ain Elzarga Wetland, why is this location one of the most important wetlands in the region?
|
The Ain Elzarga Wetland is an important wetland due to being the home for migratory waterbirds.
|
null | false
| null |
Prince of Persia is a video game franchise created by Jordan Mechner. It is built around a series of action-adventure games focused on various incarnations of the eponymous Prince, set in ancient and medieval Persia.
The first two games in the series, Prince of Persia (1989) and Prince of Persia 2: The Shadow and the Flame (1993), were published by Broderbund. Prince of Persia 3D (1999), named for being the first installment to use 3D computer graphics, was developed by Red Orb Entertainment and published by The Learning Company on PC; The Dreamcast version was developed by Avalanche Software and published by Mattel Interactive. Ubisoft bought the rights to the franchise in 2001 and rebooted it with Prince of Persia: The Sands of Time (2003). Ubisoft has since developed and published four additional entries in the series: Prince of Persia: Warrior Within (2004), Prince of Persia: The Two Thrones (2006), Prince of Persia (2008), and Prince of Persia: The Forgotten Sands (2010), as well as a number of spin-offs and games for mobile devices.
Outside of the games, the franchise also includes a film adaptation based on The Sands of Time, written in part by Mechner, and released by Walt Disney Pictures in 2010; a graphic novel; and the Lego Prince of Persia toyline. Ubisoft's Assassin's Creed franchise is considered to be the spiritual successor to the series.
|
List the name of the games in Prince of Persia series along with their publication year. List the results in comma separated format.
|
Prince of Persia (1989), Prince of Persia 2: The Shadow and the Flame (1993), Prince of Persia 3D (1999), Prince of Persia: The Sands of Time (2003), Prince of Persia: Warrior Within (2004), Prince of Persia: The Two Thrones (2006), Prince of Persia (2008), Prince of Persia: The Forgotten Sands (2010)
|
null | false
| null |
What is GeneBass?
|
GeneBass is a CRISPR-Cas9 screening platform for the high-throughput functional analysis of genes in mammalian cells. The GeneBass platform allows researchers to systematically activate or repress genes of interest and study their effects on cellular processes such as cell proliferation, differentiation, and survival. The system uses a pooled lentiviral library of guide RNAs targeting every gene in the human or mouse genome, which can be introduced into cells using viral transduction. The guide RNAs are designed to target specific regions of the genome adjacent to the transcription start sites of each gene, allowing for specific and efficient gene activation or repression. GeneBass experiments typically involve the transduction of large numbers of cells, followed by selection or sorting of cells based on specific phenotypes of interest. High-throughput sequencing of the guide RNA library before and after selection can be used to identify genes that are essential for specific cellular processes or that are involved in disease pathways.
|
|
null | false
| null |
Tell me which of these activities are exciting and which are relaxing.
|
- Reading: relaxing
- Riding on a roller coaster: exciting
- Laying in a hammock: relaxing
- Playing paintball: exciting
- Go kart racing: exciting
- Skiing: exciting
- Watching a cooking show: relaxing
- Taking a nap: relaxing
|
|
null | false
| null |
The 1972 United States presidential election in Missouri took place on November 7, 1972. All fifty states and the District of Columbia were part of the 1972 United States presidential election. Voters chose 12 electors to the Electoral College, who voted for president and vice president.
Missouri was won by the Republican nominees, incumbent President Richard Nixon of California and his running mate Vice President Spiro Agnew of Maryland. Nixon and Agnew defeated the Democratic nominees, Senator George McGovern of South Dakota and his running mate U.S. Ambassador Sargent Shriver of Maryland.
In a state substantially part of the conservative South, McGovern was viewed by many voters as a left-wing extremist because of his support for busing and civil rights, plus his opposition to the Vietnam War, support for granting amnesty to draft dodgers and support for a thousand-dollar giveaway to each American as a solution to poverty. Many, especially Republican campaigners, also believed McGovern would legalise abortion and illicit drugs if he were elected – despite the fact that his running mate Sargent Shriver was firmly opposed to abortion. These fears of McGovern's social radicalism were especially pronounced amongst poorer whites, who were abundant in the southern part of Missouri.
Nixon carried Missouri with 62.29% of the vote to McGovern's 37.71%, a victory margin of 24.58%. Nixon won all but two jurisdictions: Monroe County and St. Louis City, and was the first Republican to ever carry southern secessionist Mississippi County and Maries County, the first since Abraham Lincoln in 1864 to carry “Little Dixie” Howard and Randolph Counties, and the first since Ulysses S. Grant in 1868 to carry Lewis, Lincoln, Ralls and Ray Counties. As of the 2020 presidential election, this is the last election in which Jackson County, which contains most of Kansas City, as well as Independence, the hometown of former Democratic president Harry S. Truman (who died seven weeks after the election), voted for a Republican presidential candidate.
|
What was significant about the 1972 United States presidential election in Missouri
|
Republican candidate Richard Nixon won the presidential election in Missouri by a significant margin, nearly 25 percent, including wins in many counties that had never voted for a Republican presidential candidate before, or had rarely done so.
|
null | false
| null |
Classify the names of the Prime Ministers based on whether they belong to France or Canada.
Élisabeth Borne, Bernard Cazeneuve, Justin Trudeau, Paul Martin
|
Canada - Justin Trudeau, Paul Martin
France - Élisabeth Borne, Bernard Cazeneuve
|
|
null | false
| 32
|
A total of 90 teams registered for the shared task, and 39 of them submitted predictions for a total of 3,065 submissions. For the FLC task, 21 teams made a total of 527 submissions, and for the SLC task, 35 teams made a total of 2,538 submissions.
Below, we give an overview of the approaches as described in the participants' papers. Tables TABREF28 and TABREF29 offer a high-level summary.
A total of 90 teams registered for the shared task, and 39 of them submitted predictions for a total of 3,065 submissions.
|
How many predictions were submitted in total for the shared task?
|
3,065 submissions.
|
null | false
| 337
|
Twitter is a micro-blogging platform and a social network where users can publish and exchange short messages of up to 140 characters long (also known as tweets). Twitter has seen a great rise in popularity in recent years because of its availability and ease-of-use. This rise in popularity and the public nature of Twitter (less than 10% of Twitter accounts are private BIBREF0 ) have made it an important tool for studying the behaviour and attitude of people.
One area of research that has attracted great attention in the last few years is that of tweet sentiment classification. Through sentiment classification and analysis, one can get a picture of people's attitudes about particular topics on Twitter. This can be used for measuring people's attitudes towards brands, political candidates, and social issues.
There have been several works that do sentiment classification on Twitter using standard sentiment classification techniques, with variations of n-gram and bag of words being the most common. There have been attempts at using more advanced syntactic features as is done in sentiment classification for other domains BIBREF1 , BIBREF2 , however the 140 character limit imposed on tweets makes this hard to do as each article in the Twitter training set consists of sentences of no more than several words, many of them with irregular form BIBREF3 .
On the other hand, what tweets lack in structure they make up with sheer volume and rich metadata. This metadata includes geolocation, temporal and author information. We hypothesize that sentiment is dependent on all these contextual factors. Different locations, times and authors have different emotional valences. For instance, people are generally happier on weekends and certain hours of the day, more depressed at the end of summer holidays, and happier in certain states in the United States. Moreover, people have different baseline emotional valences from one another. These claims are supported for example by the annual Gallup poll that ranks states from most happy to least happy BIBREF4 , or the work by Csikszentmihalyi and Hunter BIBREF5 that showed reported happiness varies significantly by day of week and time of day. We believe these factors manifest themselves in sentiments expressed in tweets and that by accounting for these factors, we can improve sentiment classification on Twitter.
In this work, we explored this hypothesis by utilizing distant supervision BIBREF6 to collect millions of labelled tweets from different locations (within the USA), times of day, days of the week, months and authors. We used this data to analyse the variation of tweet sentiments across the aforementioned categories. We then used a Bayesian approach to incorporate the relationship between these factors and tweet sentiments into standard n-gram based Twitter sentiment classification.
This paper is structured as follows. In the next sections we will review related work on sentiment classification, followed by a detailed explanation of our approach and our data collection, annotation and processing efforts. After that, we describe our baseline n-gram sentiment classifier model, followed by the explanation of how the baseline model is extended to incorporate contextual information. Next, we describe our analysis of the variation of sentiment within each of the contextual categories. We then evaluate our models and finally summarize our findings and contributions and discuss possible paths for future work.
We then used a Bayesian approach to incorporate the relationship between these factors and tweet sentiments into standard n-gram based Twitter sentiment classification. This paper is structured as follows.
|
What approach does the author use to incorporate the relationship between these factors and tweet sentiments into standard n-gram-based Twitter sentiment classification?
|
A Bayesian approach.
|
null | false
| 99
|
In this work we formulate our classification problem as follows: given two classes of news articles, respectively $D$ (disinformation) and $M$ (mainstream), a set of news articles $A_i$ and associated class labels $C_i \in \lbrace D,M\rbrace $, and a set of tweets $\Pi _i=\lbrace T_i^1, T_i^2, ...\rbrace $ each of which contains an Uniform Resource Locator (URL) pointing explicitly to article $A_i$, predict the class $C_i$ of each article $A_i$. There is huge debate and controversy on a proper taxonomy of malicious and deceptive information BIBREF1BIBREF2BIBREF15BIBREF16BIBREF17BIBREF3BIBREF11. In this work we prefer the term disinformation to the more specific fake news to refer to a variety of misleading and harmful information. Therefore, we follow a source-based approach, a consolidated strategy also adopted by BIBREF6BIBREF16BIBREF2BIBREF1, in order to obtain relevant data for our analysis. We collected:
Disinformation articles, published by websites which are well-known for producing low-credibility content, false and misleading news reports as well as extreme propaganda and hoaxes and flagged as such by reputable journalists and fact-checkers;
Mainstream news, referring to traditional news outlets which deliver factual and credible information.
We believe that this is currently the most reliable classification approach, but it entails obvious limitations, as disinformation outlets may also publish true stories and likewise misinformation is sometimes reported on mainstream media. Also, given the choice of news sources, we cannot test whether our methodology is able to classify disinformation vs factual but not mainstream news which are published on niche, non-disinformation outlets.
We collected tweets associated to a dozen US mainstream news websites, i.e. most trusted sources described in BIBREF18, with the Streaming API, and we referred to Hoaxy API BIBREF16 for what concerns tweets containing links to 100+ US disinformation outlets. We filtered out articles associated to less than 50 tweets. The resulting dataset contains overall $\sim $1.7 million tweets for mainstream news, collected in a period of three weeks (February 25th, 2019-March 18th, 2019), which are associated to 6,978 news articles, and $\sim $1.6 million tweets for disinformation, collected in a period of three months (January 1st, 2019-March 18th, 2019) for sake of balance of the two classes, which hold 5,775 distinct articles. Diffusion censoring effects BIBREF14 were correctly taken into account in both collection procedures. We provide in Figure FIGREF4 the distribution of articles by source and political bias for both news domains.
As it is reported that conservatives and liberals exhibit different behaviors on online social platforms BIBREF19BIBREF20BIBREF21, we further assigned a political bias label to different US outlets (and therefore news articles) following the procedure described in BIBREF2. In order to assess the robustness of our method, we performed classification experiments by training only on left-biased (or right-biased) outlets of both disinformation and mainstream domains and testing on the entire set of sources, as well as excluding particular sources that outweigh the others in terms of samples to avoid over-fitting.
For what concerns the Italian scenario we first collected tweets with the Streaming API in a 3-week period (April 19th, 2019-May 5th, 2019), filtering those containing URLs pointing to Italian official newspapers websites as described in BIBREF22; these correspond to the list provided by the association for the verification of newspaper circulation in Italy (Accertamenti Diffusione Stampa). We instead referred to the dataset provided by BIBREF23 to obtain a set of tweets, collected continuously since January 2019 using the same Twitter endpoint, which contain URLs to 60+ Italian disinformation websites. In order to get balanced classes (April 5th, 2019-May 5th, 2019), we retained data collected in a longer period w.r.t to mainstream news. In both cases we filtered out articles with less than 50 tweets; overall this dataset contains $\sim $160k mainstream tweets, corresponding to 227 news articles, and $\sim $100k disinformation tweets, corresponding to 237 news articles. We provide in Figure FIGREF5 the distribution of articles according to distinct sources for both news domains. As in the US dataset, we took into account censoring effects BIBREF14 by excluding tweets published before (left-censoring) or after two weeks (right-censoring) from the beginning of the collection process.
The different volumes of news shared on Twitter in the two countries are due both to the different population size of US and Italy (320 vs 60 millions) but also to the different usage of Twitter platform (and social media in general) for news consumption BIBREF24. Both datasets analyzed in this work are available from the authors on request.
A crucial aspect in our approach is the capability to fully capturing sharing cascades on Twitter associated to news articles. It has been reported BIBREF25 that the Twitter streaming endpoint filters out tweets matching a given query if they exceed 1% of the global daily volume of shared tweets, which nowadays is approximately $5\cdot 10^8$; however, as we always collected less than $10^6$ tweets per day, we did not incur in this issue and we thus gathered 100% of tweets matching our query.
We built Twitter diffusion networks following an approach widely adopted in the literature BIBREF6BIBREF17BIBREF2. We remark that there is an unavoidable limitation in Twitter Streaming API, which does not allow to retrieve true re-tweeting cascades because re-tweets always point to the original source and not to intermediate re-tweeting users BIBREF8BIBREF14; thus we adopt the only viable approach based on Twitter's public availability of data. Besides, by disentangling different interactions with multiple layers we potentially reduce the impact of this limitation on the global network properties compared to the single-layer approach used in our baseline.
Using the notation described in BIBREF26. we employ a multi-layer representation for Twitter diffusion networks. Sociologists have indeed recognized decades ago that it is crucial to study social systems by constructing multiple social networks where different types of ties among same individuals are used BIBREF27. Therefore, for each news article we built a multi-layer diffusion network composed of four different layers, one for each type of social interaction on Twitter platform, namely retweet (RT), reply (R), quote (Q) and mention (M), as shown in Figure FIGREF11. These networks are not necessarily node-aligned, i.e. users might be missing in some layers. We do not insert "dummy" nodes to represent all users as it would have severe impact on the global network properties (e.g. number of weakly connected components). Alternatively one may look at each multi-layer diffusion network as an ensemble of individual graphs BIBREF26; since global network properties are computed separately for each layer, they are not affected by the presence of any inter-layer edges.
In our multi-layer representation, each layer is a directed graph where we add edges and nodes for each tweet of the layer type, e.g. for the RT layer: whenever user $a$ retweets account $b$ we first add nodes $a$ and $b$ if not already present in the RT layer, then we build an edge that goes from $b$ to $a$ if it does not exists or we increment the weight by 1. Similarly for the other layers: for the R layer edges go from user $a$ (who replies) to user $b$, for the Q layer edges go from user $b$ (who is quoted by) to user $a$ and for the M layer edges go from user $a$ (who mentions) to user $b$.
Note that, by construction, our layers do not include isolated nodes; they correspond to "pure tweets", i.e. tweets which have not originated any interactions with other users. However, they are present in our dataset, and their number is exploited for classification, as described below.
We used a set of global network indicators which allow us to encode each network layer by a tuple of features. Then we simply concatenated tuples as to represent each multi-layer network with a single feature vector. We used the following global network properties:
Number of Strongly Connected Components (SCC): a Strongly Connected Component of a directed graph is a maximal (sub)graph where for each pair of vertices $u,v$ there is a path in each direction ($u\rightarrow v$, $v\rightarrow u$).
Size of the Largest Strongly Connected Component (LSCC): the number of nodes in the largest strongly connected component of a given graph.
Number of Weakly Connected Components (WCC): a Weakly Connected Component of a directed graph is a maximal (sub)graph where for each pair of vertices $(u, v)$ there is a path $u \leftrightarrow v$ ignoring edge directions.
Size of the Largest Weakly Connected Component (LWCC): the number of nodes in the largest weakly connected component of a given graph.
Diameter of the Largest Weakly Connected Component (DWCC): the largest distance (length of the shortest path) between two nodes in the (undirected version of) largest weakly connected component of a graph.
Average Clustering Coefficient (CC): the average of the local clustering coefficients of all nodes in a graph; the local clustering coefficient of a node quantifies how close its neighbours are to being a complete graph (or a clique). It is computed according to BIBREF28.
Main K-core Number (KC): a K-core BIBREF13 of a graph is a maximal sub-graph that contains nodes of internal degree $k$ or more; the main K-core number is the highest value of $k$ (in directed graphs the total degree is considered).
Density (d): the density for directed graphs is $d=\frac{|E|}{|V||V-1|}$, where $|E|$ is the number of edges and $|N|$ is the number of vertices in the graph; the density equals 0 for a graph without edges and 1 for a complete graph.
Structural virality of the largest weakly connected component (SV): this measure is defined in BIBREF14 as the average distance between all pairs of nodes in a cascade tree or, equivalently, as the average depth of nodes, averaged over all nodes in turn acting as a root; for $|V| > 1$ vertices, $SV=\frac{1}{|V||V-1|}\sum _i\sum _j d_{ij}$ where $d_{ij}$ denotes the length of the shortest path between nodes $i$ and $j$. This is equivalent to compute the Wiener's index BIBREF29 of the graph and multiply it by a factor $\frac{1}{|V||V-1|}$. In our case we computed it for the undirected equivalent graph of the largest weakly connected component, setting it to 0 whenever $V=1$.
We used networkx Python package BIBREF30 to compute all features. Whenever a layer is empty. we simply set to 0 all its features. In addition to computing the above nine features for each layer, we added two indicators for encoding information about pure tweets, namely the number T of pure tweets (containing URLs to a given news article) and the number U of unique users authoring those tweets. Therefore, a single diffusion network is represented by a vector with $9\cdot 4+2=38$ entries.
Aforementioned network properties can be qualitatively explained in terms of social footprints as follows: SCC correlates with the size of the diffusion network, as the propagation of news occurs in a broadcast manner most of the time, i.e. re-tweets dominate on other interactions, while LSCC allows to distinguish cases where such mono-directionality is somehow broken. WCC equals (approximately) the number of distinct diffusion cascades pertaining to each news article, with exceptions corresponding to those cases where some cascades merge together via Twitter interactions such as mentions, quotes and replies, and accordingly LWCC and DWCC equals the size and the depth of the largest cascade. CC corresponds to the level of connectedness of neighboring users in a given diffusion network whereas KC identifies the set of most influential users in a network and describes the efficiency of information spreading BIBREF17. Finally, d describes the proportions of potential connections between users which are actually activated and SV indicates whether a news item has gained popularity with a single and large broadcast or in a more viral fashion through multiple generations.
For what concerns different Twitter actions, users primarily interact with each other using retweets and mentions BIBREF20.
The former are the main engagement activity and act as a form of endorsement, allowing users to rebroadcast content generated by other users BIBREF31. Besides, when node B retweets node A we have an implicit confirmation that information from A appeared in B's Twitter feed BIBREF12. Quotes are simply a special case of retweets with comments.
Mentions usually include personal conversations as they allow someone to address a specific user or to refer to an individual in the third person; in the first case they are located at the beginning of a tweet and they are known as replies, otherwise they are put in the body of a tweet BIBREF20. The network of mentions is usually seen as a stronger version of interactions between Twitter users, compared to the traditional graph of follower/following relationships BIBREF32.
We collected tweets associated to a dozen US mainstream news websites, i.e. most trusted sources described in (Mitchell et al. 2014), with the Streaming API, and we re_x0002_ferred to Hoaxy API (Shao et al. 2016) for what concerns tweets containing links to 100+ US disinformation outlets.
|
How to collect tweets by the authors?
|
The authors collected tweets associated to a dozen US mainstream news websites with the Streaming API and referred to Hoaxy API for what concerns tweets containing links to 100+ US disinformation outlets.
|
null | false
| null |
The Sun Also Rises is a 1926 novel by American writer Ernest Hemingway, his first, that portrays American and British expatriates who travel from Paris to the Festival of San Fermín in Pamplona to watch the running of the bulls and the bullfights. An early modernist novel, it received mixed reviews upon publication. Hemingway biographer Jeffrey Meyers writes that it is now "recognized as Hemingway's greatest work" and Hemingway scholar Linda Wagner-Martin calls it his most important novel. The novel was published in the United States in October 1926 by Scribner's. A year later, Jonathan Cape published the novel in London under the title Fiesta. It remains in print.
The novel is a roman à clef: the characters are based on people in Hemingway's circle and the action is based on events, particularly Hemingway's life in Paris in the 1920s and a trip to Spain in 1925 for the Pamplona festival and fishing in the Pyrenees. Hemingway presents his notion that the "Lost Generation"—considered to have been decadent, dissolute and irretrievably damaged by World War I—was in fact resilient and strong. Hemingway investigates the themes of love and death, the revivifying power of nature and the concept of masculinity. His spare writing style, combined with his restrained use of description to convey characterizations and action, demonstrates his "Iceberg Theory" of writing.
|
What is the book "The Sun Also Rises" about?
|
The Sun Also Rises is a book based on people Hemingway knew from his life in Paris, fishing in the Pyrenees, and a trip to the Pamplona festival in Spain.
|
null | false
| null |
Why do we have fences?
|
Fences are used to physically show where your home lot start and where it ends. It is also used to give you privacy from your neighbors and private space you can enjoy with friends, family and pets. Fences can be decorated with plants or lights, or they can be painted to match another exterior color (e.g. house bricks). Fences are made out wood planks that are nailed to perpendicular 2.x4 planks of wood with steel nails.
|
|
null | false
| 221
|
Spoken language understanding (SLU) systems infer the meaning or intent of a spoken utterance BIBREF0 . This is crucial for voice user interfaces, in which the speaker's utterance needs to be converted into an action or query. For example, for a voice-controlled coffee machine, an utterance like “make me a large coffee with two milks and a sugar, please” might have an intent representation like {drink: "coffee", size: "large", additions: [{type: "milk", count: 2}, {type: "sugar", count: 1}]}.
The conventional SLU pipeline is composed of two modules: an automatic speech recognition (ASR) module that maps the speech to a text transcript, and a natural language understanding (NLU) module that maps the text transcript to the speaker's intent BIBREF1 , BIBREF2 , BIBREF3 . An alternative approach that is beginning to gain popularity is end-to-end SLU BIBREF4 , BIBREF5 , BIBREF6 , BIBREF7 , BIBREF8 . In end-to-end SLU, a single trainable model maps the speech audio directly to the speaker's intent without explicitly producing a text transcript (Fig. FIGREF4 ). Unlike the conventional SLU pipeline, end-to-end SLU:
End-to-end models have been made possible by deep learning, which automatically learns hierarchical representations of the input signal BIBREF9 , BIBREF10 , BIBREF11 , BIBREF12 , BIBREF13 . Speech is natural to represent in a hierarchical way: waveform INLINEFORM0 phonemes INLINEFORM1 morphemes INLINEFORM2 words INLINEFORM3 concepts INLINEFORM4 meaning. However, because speech signals are high-dimensional and highly variable even for a single speaker, training deep models and learning these hierarchical representations without a large amount of training data is difficult.
The computer vision BIBREF14 , BIBREF15 , natural language processing BIBREF16 , BIBREF17 , BIBREF18 , BIBREF19 , BIBREF20 , and ASR BIBREF21 , BIBREF22 communities have attacked the problem of limited supervised training data with great success by pre-training deep models on related tasks for which there is more training data. Following their lead, we propose an efficient ASR-based pre-training methodology in this paper and show that it may be used to improve the performance of end-to-end SLU models, especially when the amount of training data is very small.
Our contributions are as follows:
Following their lead, we propose an efficient ASR-based pre-training methodology in this paper and show that it may be used to improve the performance of end-to-end SLU models, especially when the amount of training data is very small.
|
Why do they propose an efficient automatic speech recognition (ASR)-based pre-training methodology?
|
To improve the performance of end-to-end Spoken language understanding (SLU) models, especially when the amount of training data is very small.
|
1806.09103
| false
| null |
In our experiments, the short list is determined according to the word frequency. Concretely, we sort the vocabulary according to the word frequency from high to low. A frequency filter ratio INLINEFORM0 is set to filter out the low-frequency words (rare words) from the lookup table. For example, INLINEFORM1 =0.9 means the least frequent 10% words are replaced with the default UNK notation.
A frequency filter ratio INLINEFORM0 is set to filter out the low-frequency words (rare words) from the lookup table
|
how are rare words defined?
|
The answers are shown as follows:
* low-frequency words
|
null | false
| 353
|
Political event data has existed in various forms since the 1970s. Two of the most common political event datasets were the World Event Interaction Survey (WEIS) and the Conflict and Peace Data Bank (COPDAB) BIBREF0 , BIBREF1 . These two datasets were eventually replaced by the projects created by Philip Schrodt and various collaborators. In general, these projects were marked by the use of the Conflict and Mediation Event Observations (CAMEO) coding ontology and automated, machine-coding rather than human coding BIBREF2 , BIBREF3 . The CAMEO ontology is made up of 20 “top-level” categories that encompass actions such as “Make Statement” or “Protest”, and contains over 200 total event classifications. This ontology has served as the basis for most of the modern event datasets such as the Integrated Crisis Early Warning System (ICEWS) BIBREF4 , the Global Database of Events, Language, and Tone (GDELT), and the Phoenix dataset presented in this paper.
This type of data can prove highly useful for many types of studies. Since this type of data is inherently atomic, each observation is a record of a single event between a source and a target, it provides a disaggregated view of political events. This means that the data can be used to examine interactions below the usual monthly or yearly levels of aggregation. This approach can be used in a manner consistent with traditional hypothesis testing that is the norm in political science BIBREF5 , BIBREF6 , BIBREF7 . Additionally, event data has proven useful in forecasting models of conflict since the finer time resolution allows analysts to gain better leverage over the prediction problem than is possible when using more highly aggregated data BIBREF8 , BIBREF9 , BIBREF10 , BIBREF11 . Finally, the advent of daily-updated event data has led to many novel uses such as watchboarding or dashboarding. The goal in these situations is to provide an easy to understand interface that analysts can use to quickly monitor ongoing or emerging situations around the world. These applications provide a new frontier for event data that has not been considered much until this point.
The status quo of TABARI-generated, CAMEO-coded event data, which was established in the early 2000s, has remained with little change. BIBREF12 outlined many potential advances in the generation of political event data. These advances are things such as realtime processing of news stories, the incorporation of open-source natural language processing (NLP) software, and enhancements in the automated coding structure. Two publicly-available datasets, GDELT and ICEWS, have each attempted to implement some, or all, of these changes in their respective data-generating pipelines. In terms of goals, the ICEWS project seems closest to sharing the vision of the Phoenix dataset. A more in-depth comparison of Phoenix and ICEWS is presented in a later section. In short, the goal of the project presented in this chapter is to implement most of the improvements suggested in BIBREF12 .
Political event data has existed in various forms since the 1970s. Two of the most common political event datasets were the World Event Interaction Survey (WEIS) and the Conflict and Peace Data Bank (COPDAB) (Azar 1980; McClelland 1976).
|
What were the two of the most common political event datasets
|
The World Event Interaction Survey (WEIS) and the Conflict and Peace Data Bank (COPDAB).
|
null | false
| 339
|
For an ever increasing spectrum of applications (e.g., medical text analysis, opinion mining, sentiment analysis, social media text analysis, customer intelligence, fraud analytics etc.) mining and analysis of unstructured natural language text data is necessary BIBREF0, BIBREF1, BIBREF2.
One of key challenge while designing such text analytics (TA) applications is to identify right set of features. For example, for text classification problem, different sets of features have been considered in different works (spanning a history of more than twenty years) including `bag of words', `bag of phrases', `bag of n-grams', `WordNet based word generalizations', and `word embeddings' BIBREF3, BIBREF4, BIBREF5, BIBREF6, BIBREF7. Even for recent end-to-end designs using deep neural networks, specification of core features remains manually driven BIBREF8, BIBREF9. During feature engineering, often data scientists manually determine which features to use based upon their experience and expertise with respect to the underlying application domain as well as state-of-the-art tools and techniques. Different tools (e.g., NLTK BIBREF10, Mallet BIBREF11, Stanford CoreNLP BIBREF12, Apache OpenNLP BIBREF13, Apache Lucene BIBREF14, etc.) available to a NLP data scientist for TA application design and development often differ in terms of support towards extraction of features, level of granularity at which feature extraction process is to be specified; and these tools often use different programing vocabularies to specify semantically equivalent features.
Currently, there is no generic method or approach, which can be applied during TA application's design process to define and extract features for any arbitrary application in an automated or semi-automated manner. Even there is no single way to express wide range of NLP features, resulting into increased efforts during feature engineering which has to start new for each data scientist and automated reuse of features across semantically similar or related applications designed by different data scientists is difficult. This also hinders foundational studies on NLP feature engineering including why certain features are more critical than others.
In this paper, we aim to present an approach towards automating NLP feature engineering. We start with an outline of a language for expressing NLP features abstracting over the feature extraction process, which often implicitly captures intent of the NLP data scientist to extract specific features from given input text. We next discuss a method to enable automated reuse of features across semantically related applications when a corpus of feature specifications for related applications is available. Proposed language and system would help achieving reduction in manual effort towards design and extraction of features, would ensure standardization in feature specification process, and could enable effective reuse of features across similar and/or related applications.
Proposed language and system would help achieving reduction in manual effort towards design and extraction of features, would ensure standardization in feature specification process, and could enable effective reuse of features across similar and/or related applications.
|
What is the use of the proposed language?
|
The proposed language would help achieving reduction in manual effort towards design and extraction of features, would ensure standardization in feature specification process, and could enable effective reuse of features across similar and/or related applications.
|
null | false
| null |
What is 12-3-30?
|
12-3-3- is the viral workout trend that's been popularized on social media. It entails walking on a treadmill at an incline of 12, a speed of 3 mph, for 30 minutes.
|
|
1909.13362
| false
| null |
FLOAT SELECTED: TABLE II REPORTED ACCURACIES OF STATE OF THE ART AND SELECTED HIGH PERFORMING SYLLABIFIERS ON EACH EVALUATION DATASET.
For each dataset used to evaluate the proposed model, we compare our results with published accuracies of existing syllabification systems. Table TABREF21 shows the performance of well known and state of the art syllabifiers for each dataset. Liang's hyphenation algorithm is commonly known for its usage in . The patgen program was used to learn the rules of syllable boundaries BIBREF39. What we call Entropy CRF is a method particular to Manipuri; a rule-based component estimates the entropy of phones and phone clusters while a data-driven CRF component treats syllabification as a sequence modeling task BIBREF35.
FLOAT SELECTED: TABLE II REPORTED ACCURACIES OF STATE OF THE ART AND SELECTED HIGH PERFORMING SYLLABIFIERS ON EACH EVALUATION DATASET.
For each dataset used to evaluate the proposed model, we compare our results with published accuracies of existing syllabification systems. Table TABREF21 shows the performance of well known and state of the art syllabifiers for each dataset.
|
Which models achieve state-of-the-art performances?
|
CELEX (Dutch and English) - SVM-HMM
Festival, E-Hitz and OpenLexique - Liang hyphenation
IIT-Guwahat - Entropy CRF
|
null | false
| null |
East Side of Manhattan refers to the side of Manhattan which abuts the East River and faces Brooklyn and Queens. Fifth Avenue, Central Park from 59th to 110th Streets, and Broadway below 8th Street separate it from the West Side.
The major neighborhoods on the East Side include (from north to south) East Harlem, Yorkville, the Upper East Side, Turtle Bay, Murray Hill, Kips Bay, Gramercy, East Village, and the Lower East Side. The main north–south expressways servicing the East Side are the Franklin D. Roosevelt East River Drive and Harlem River Drive, which for the majority of their length are separated from the east shore of the island by the Manhattan Waterfront Greenway. The East Side is served by the IRT East Side Line subway and by many bus lines.
The West Side of Manhattan refers to the side of Manhattan Island which abuts the Hudson River and faces the U.S. state of New Jersey. Fifth Avenue, Central Park, and lower Broadway separate it from the East Side. The major neighborhoods on the West Side are (from north to south) Inwood, Hudson Heights, Washington Heights, West Harlem, Morningside Heights, Manhattan Valley, Upper West Side, Hell's Kitchen, Chelsea, West Village, SoHo, and Tribeca. The 8th Avenue and West Side subway lines connect all parts of the West Side.
The main north-south roads servicing the West Side are the Henry Hudson Parkway in the north, and the West Side Highway in the south. The Hudson River Greenway separates those highways from the western shore of the island.
|
List the major neighborhoods on the East Side and West Side of Manhattan.
|
East Side: East Harlem, Yorkville, the Upper East Side, Turtle Bay, Murray Hill, Kips Bay, Gramercy, East Village, and the Lower East Side
West Side: Inwood, Hudson Heights, Washington Heights, West Harlem, Morningside Heights, Manhattan Valley, Upper West Side, Hell's Kitchen, Chelsea, West Village, SoHo, and Tribeca
|
null | false
| null |
Thomas Cruise Mapother IV (born July 3, 1962) is an American actor and producer. One of the world's highest-paid actors, he has received various accolades, including an Honorary Palme d'Or and three Golden Globe Awards, in addition to nominations for four Academy Awards. His films have grossed over $4 billion in North America and over $11.5 billion worldwide, making him one of the highest-grossing box-office stars of all time.
|
Given a reference text about Tom Cruise, tell me when they were born and what they are famous for.
|
Tom Cruise was born on July 3rd, 1962. He is famous for being an American actor and producer and winning three golden globes.
|
null | false
| null |
When should I use Slack versus email versus a phone call?
|
For quick communication and memes, Slack is a good medium. For longer form communication, or those that include external parties that are not available on Slack, email is an excellent choice. Phone calls are useful for instantaneous feedback, fluid dialog, and establishing emotional rapport with the participants.
|
|
null | false
| null |
What are some unique and interesting travel destinations in the UK?
|
Here are some of the most interesting travel destinations in the UK?
- London
- Edinburgh
- Stonehenge
- The Cotswolds
- Cambridge
- Oxford
- Liverpool
- Loch Ness
- Giant's Causeway
- Belfast
- Cardiff
|
|
null | false
| null |
Classify these animals by either a mammal or reptile;
Lizard, Alligator, Elephant, Crocodile, Lion, Dolphin, Snake, Platypus, Chimpanzee
Turtle
|
Yes, I grouped these animals as mammals and reptiles. Here is the categories;
Mammals: Lion, Elephant, Dolphin, Platypus, Chimpanzee
Reptiles: Turtle, Crocodile, Snake, Lizard, Alligator
|
|
null | false
| null |
Why do people hate pineapple on pizza?
|
Pineapple has grown to be a rather controversial pizza topping. Those that love it may like it for the contrast of the sweet and juicy pineapple combined with the savory profile from a traditional tomato sauce based pizza. As pineapple is commonly known to be from Hawaii, some pizzerias have named a pineapple and ham pizza as a "Hawaiian" pizza which some may favor as Hawaii is usually a popular vacation destination and considered an exotic option for something traditional like pizza. Others that take a more purist preference towards pizza toppings have criticized pineapple as an abomination.
|
|
null | false
| null |
Bowling achievements of Andrew Kyle
|
Andrew Kyle (born 1978) is a Northern Irish international lawn and indoor bowler.
Kyle bowls for County Antrim Bowling Club indoors and Larne Bowling Club outdoors and in 2016 represented a combined Irish team at the Bowls World Cup in Australia. In 2017, he won three medals at the European Bowls Championships.
He was selected as part of the Northern Ireland team for the 2018 Commonwealth Games on the Gold Coast in Queensland.
In 2019, he won the fours bronze medal at the Atlantic Bowls Championships.
|
|
1912.00159
| false
| null |
Table TABREF32 shows the perplexity of the models on each of the test sets. As expected, each model performs better on the test set they have been trained on. When applied to a different test set, both see an increase in perplexity. However, the Leipzig model seems to have more trouble generalizing: its perplexity nearly doubles on the SwissCrawl test set and raises by twenty on the combined test set.
Table TABREF32 shows the perplexity of the models on each of the test sets. As expected, each model performs better on the test set they have been trained on. When applied to a different test set, both see an increase in perplexity. However, the Leipzig model seems to have more trouble generalizing: its perplexity nearly doubles on the SwissCrawl test set and raises by twenty on the combined test set.
|
How is language modelling evaluated?
|
The answers are shown as follows:
* perplexity of the models
|
null | false
| null |
Are these blue or red colors? Pink, navy, maroon, mahogany, teal, sky, robin, scarlet
|
Pink: red, navy: blue, maroon: red, mahogany: red, teal: blue, sky: blue, robin: blue, scarlet: red
|
|
null | false
| 147
|
CLIR systems retrieve documents written in a language that is different from search query language BIBREF0 . The primary objective of CLIR is to translate or project a query into the language of the document repository BIBREF1 , which we refer to as Retrieval Corpus (RC). To this end, common CLIR approaches translate search queries using a Machine Translation (MT) model and then use a monolingual IR system to retrieve from RC. In this process, a translation model is treated as a black box BIBREF2 , and it is usually trained on a sentence level parallel corpus, which we refer to as Translation Corpus (TC).
We address a pitfall of using existing MT models for query translation BIBREF1 . An MT model trained on TC does not have any knowledge of RC. In an extreme setting, where there are no common terms between the target side of TC and RC, a well trained and tested translation model would fail because of vocabulary mismatch between the translated query and documents of RC. Assuming a relaxed scenario where some commonality exists between two corpora, a translation model might still perform poorly, favoring terms that are more likely in TC but rare in RC. Our hypothesis is that a search query translation model would perform better if a translated query term is likely to appear in the both retrieval and translation corpora, a property we call balanced translation.
To achieve balanced translations, it is desired to construct an MT model that is aware of RC vocabulary. Different types of MT approaches have been adopted for CLIR task, such as dictionary-based MT, rule-based MT, statistical MT etc. BIBREF3 . However, to the best of our knowledge, a neural search query translation approach has yet to be taken by the community. NMT models with attention based encoder-decoder techniques have achieved state-of-the-art performance for several language pairs BIBREF4 . We propose a multi-task learning NMT architecture that takes RC vocabulary into account by learning Relevance-based Auxiliary Task (RAT). RAT is inspired from two word embedding learning approaches: Relevance-based Word Embedding (RWE) BIBREF5 and Continuous Bag of Words (CBOW) embedding BIBREF6 . We show that learning NMT with RAT enables it to generate balanced translation.
NMT models learn to encode the meaning of a source sentence and decode the meaning to generate words in a target language BIBREF7 . In the proposed multi-task learning model, RAT shares the decoder embedding and final representation layer with NMT. Our architecture answers the following question: In the decoding stage, can we restrict an NMT model so that it does not only generate terms that are highly likely in TC?. We show that training a strong baseline NMT with RAT roughly achieves 16% improvement over the baseline. Using a qualitative analysis, we further show that RAT works as a regularizer and prohibits NMT to overfit to TC vocabulary.
The primary objective of CLIR is to translate or project a query into the language of the document repository (Sokokov et al., 2013), which we refer to as Retrieval Corpus (RC).
|
What is the primary objective of CLIR?
|
The primary objective of CLIR is to translate or project a query into the language of the document repository , which the authors refer to as Retrieval Corpus (RC).
|
null | false
| null |
What is an Ironman Triathlon?
|
An Ironman Triathlon is a multi-sport event that consists of a 2.4 mile swim, 112 mile bike ride, and 26.2 mile run.
|
|
1909.00088
| true
| null |
We propose a pipeline called SMERTI (pronounced `smarty') for STE. Combining entity replacement (ER), similarity masking (SM), and text infilling (TI), SMERTI can modify the semantic content of text. We define a metric called the Semantic Text Exchange Score (STES) that evaluates the overall ability of a model to perform STE, and an adjustable parameter masking (replacement) rate threshold (MRT/RRT) that can be used to control the amount of semantic change.
We define a metric called the Semantic Text Exchange Score (STES) that evaluates the overall ability of a model to perform STE, and an adjustable parameter masking (replacement) rate threshold (MRT/RRT) that can be used to control the amount of semantic change.
|
Has STES been previously used in the literature to evaluate similar tasks?
|
No.
|
null | false
| null |
Identify which animal species is alive or extinct: Mosasaurus, Whale Shark
|
Whale Shark is alive, Mosasaurus is extinct.
|
|
null | false
| null |
Identify which car manufacturer is Japanese or American: RAM, Toyota Motor Corporation
|
RAM is American, Toyota Motor Corporation is Japanese
|
|
null | false
| 356
|
In this section, we provide comparative results on our new Multilingual Document Classification Corpus. Since the initial work by BIBREF0 many alternative approaches to cross-lingual document classification have been developed. We will encourage the respective authors to evaluate their systems on MLDoc. We believe that a large variety of transfer language pairs will give valuable insights on the performance of the various approaches.
In this paper, we propose initial strong baselines which represent two complementary directions of research: one based on the aggregation of multilingual word embeddings, and another one, which directly learns multilingual sentence representations. Details on each approach are given in section "Multilingual word representations" and "Multilingual sentence representations" respectively. In contrast to previous works on cross-lingual document classification with RVC2, we explore training the classifier on all languages and transfer it to all others, ie. we do not limit our study to the transfer between English and a foreign language.
One can envision several ways to define cross-lingual document classification, in function of the resources which are used in the source and transfer language (see Table 3 ). The first scheme assumes that we have no resources in the transfer language at all, neither labeled nor unlabeled. We will name this case “zero-shot cross-lingual document classification”. To simplify the presentation, we will assume that we transfer from English to German. The training and evaluation protocol is as follows. First, train a classifier using resources in the source language only, eg. the training and development corpus are in English. All meta parameters and model choices are performed using the English development corpus. Once the best performing model is selected, it is applied to the transfer language, eg. the German test set. Since no resources of the transfer language are used, the same system can be applied to many different transfer languages. This type of cross-lingual document classification needs a very strong multilingual representation since no knowledge on the target language was used during the development of the classifier.
In a second class of cross-lingual document classification, we may aim in improving the transfer performance by using a limited amount of resources in the target language. In the framework of the proposed MLDoc we will use the development corpus of target language for model selection. We will name this method “targeted cross-lingual document classification” since the system is tailored to one particular transfer language. It is unlikely that this system will perform well on other languages than the ones used for training or model selection.
If the goal is to build one document classification system for many languages, it may be interesting to use already several languages during training and model selection. To allow a fair comparison, we will assume that these multilingual resources have the same size than the ones used for zero-shot or targeted cross-language document classification, e.g. a training set composed of five languages with 200 examples each. This type of training is not a cross-lingual approach any more. Consequently, we will refer to this method as “joint multilingual document classification”.
We will name this method “targeted cross-lingual document classification” since the system is tailored to one particular transfer language.
|
Why do the authors name the method as "targeted cross-lingual document classification"?
|
Because the system is tailored to one particular transfer language.
|
null | false
| 388
|
In the context of investigative journalism, we address the problem of automatically identifying which claims in a given document are most worthy and should be prioritized for fact-checking. Despite its importance, this is a relatively understudied problem. Thus, we create a new dataset of political debates, containing statements that have been fact-checked by nine reputable sources, and we train machine learning models to predict which claims should be prioritized for fact-checking, i.e., we model the problem as a ranking task. Unlike previous work, which has looked primarily at sentences in isolation, in this paper we focus on a rich input representation modeling the context: relationship between the target statement and the larger context of the debate, interaction between the opponents, and reaction by the moderator and by the public. Our experiments show state-of-the-art results, outperforming a strong rivaling system by a margin, while also confirming the importance of the contextual information.
Thus, we create a new dataset of political debates, containing statements that have been fact-checked by nine reputable sources, and we train machine learning models to predict which claims should be prioritized for fact-checking, i.e., we model the problem as a ranking task.
|
How to model the problem?
|
The authors train machine learning models to predict which claims should be prioritized for fact-checking, i.e., they model the problem as a ranking task.
|
null | false
| null |
Create a checklist to move from a 3 bedroom apartment to a 5 bedroom home
|
1. Declutter first
2. Pack one room at a time and label the boxes
3. Mark Breakables
4. Create a first aid box
5. Create an essentials box
6. Move boxes into the respective rooms
7. Unpack
8. Re-arrange
9. Trash packaging and donate boxes
|
|
null | false
| 209
|
Language variety identification aims at labelling texts in a native language (e.g. Spanish, Portuguese, English) with their specific variation (e.g. Argentina, Chile, Mexico, Peru, Spain; Brazil, Portugal; UK, US). Although at first sight language variety identification may seem a classical text classification problem, cultural idiosyncrasies may influence the way users construct their discourse, the kind of sentences they build, the expressions they use or their particular choice of words. Due to that, we can consider language variety identification as a double problem of text classification and author profiling, where information about how language is shared by people may help to discriminate among classes of authors depending on their language variety.
This task is specially important in social media. Despite the vastness and accessibility of the Internet destroyed frontiers among regions or traits, companies are still very interested in author profiling segmentation. For example, when a new product is launched to the market, knowing the geographical distribution of opinions may help to improve marketing campaigns. Or given a security threat, knowing the possible cultural idiosyncrasies of the author may help to better understand who could have written the message.
Language variety identification is a popular research topic of natural language processing. In the last years, several tasks and workshops have been organized: the Workshop on Language Technology for Closely Related Languages and Language Variants @ EMNLP 2014; the VarDial Workshop @ COLING 2014 - Applying NLP Tools to Similar Languages, Varieties and Dialects; and the LT4VarDial - Joint Workshop on Language Technology for Closely Related Languages, Varieties and Dialect @ RANLP BIBREF0 BIBREF1 . We can find also several works focused on the task. In BIBREF2 the authors addressed the problem of identifying Arabic varieties in blogs and social fora. They used character $n$ -gram features to discriminate between six different varieties and obtained accuracies between 70%-80%. Similarly, BIBREF3 collected 1,000 news articles of two varieties of Portuguese. They applied different features such as word and character $n$ -grams and reported accuracies over 90%. With respect to the Spanish language, BIBREF4 focused on varieties from Argentina, Chile, Colombia, Mexico and Spain in Twitter. They used meta-learning and combined four types of features: i) character $n$ -gram frequency profiles, ii) character $n$ -gram language models, iii) Lempel-Ziv-Welch compression and iv) syllable-based language models. They obtained an interesting 60%-70% accuracy of classification.
We are interested in discovering which kind of features capture higher differences among varieties. Our hypothesis is that language varieties differ mainly in lexicographic clues. We show an example in Table 1 .
In this work we focus on the Spanish language variety identification. We differentiate from the previous works as follows: i) instead of $n$ -gram based representations, we propose a low dimensionality representation that is helpful when dealing with big data in social media; ii) in order to reduce the possible over-fitting, our training and test partitions do not share any author of instance between them; and iii) in contrast to the Twitter dataset of BIBREF4 , we will make available our dataset to the research community.
We are interested in discovering which kind of features capture higher differences among varieties. Our hypothesis is that language varieties differ mainly in lexicographic clues.
|
What is the authors' hypothesis in this paper?
|
The hypothesis is that language varieties differ mainly in lexicographic clues.
|
null | false
| null |
Imagine you are getting married and are attending a wedding cake tasting. Describe five different flavors you try, with combinations of both cake and frosting for each tasting. Pick one best flavor at the end of the five tastings.
|
The first flavor I try is vanilla cake with a buttercream frosting. The cake is good, but too plain and basic. Next we try a vanilla cake with a chai frosting, and I like it a lot more, since I love spice. Although I do worry that the vanilla cake might be a bit plain. A third option is a dark chocolate cake with a cherry and chocolate ganache filling, which is very rich and moist, which I appreciate. Fourth we try a lemon drizzle cake, which is nice but I'm not sure the right choice for the rest of our menu. And for our final flavor we try a matcha cake with white chocolate frosting, which has a very nice cake but I think the frosting is too sweet. Finally, I decide to pick the chocolate and cherry combination because I think it had the best cake and frosting combination.
|
|
null | false
| null |
While deciding to buy a bicycle, you need to consider few parameters such as age, gender, price, quality, the purpose of buying, type of bicycle/features etc. For ex: if you are a boy of age 24, then you would consider to buy a bicycle with height of 18" inch, with gear & shocker, etc.
And if you are a girl of age 6 yrs, then you would consider a kid friendly bicycle with good safety features that the kid can ride.
|
Budget, quality and the person who will ride the bicycle are key thing. Start with exploring the options based on budget and the type of bicycle that you may need and then narrow down the bicycle basis your requirement. And then compare with features, price and quality that can help to narrow down the selection process.
|
|
null | false
| null |
The corridor is used by many Amtrak trains, including the high-speed Acela, intercity trains and several long-distance trains. Most of the corridor also has frequent commuter rail service, operated by the MBTA, Shore Line East, Hartford Line, Metro-North Railroad, Long Island Rail Road, New Jersey Transit, SEPTA and MARC. While large through freights have not run on the NEC since the early 1980s, several companies continue to run smaller local freights over some select few sections of the NEC including CSX, Norfolk Southern, CSAO, Providence and Worcester, New York and Atlantic and Canadian Pacific, with the first two considered to have part-ownership over those routes.
|
From the passage provided, extract the freight train companies using the corridor. Place them in a bullet list.
|
- CSX
- Norfolk Southern
- CSAO
- Providence and Worcester
- New York and Atlantic
- Canadian Pacific
|
null | false
| null |
What are 5 things to do when visiting Oahu, Hawaii?
|
There are lots of things to do when visiting the Island of Oahu, Hawaii. Here are 5 common activities:
1. Go snorkeling and experience some of Hawaii’s amazing marine life at Hanauma Bay.
2. Walk the streets of downtown Honolulu and enjoy Hawaiian shave ice.
3. Take a trip to the Dole plantation to learn more about pineapples and try some of the famous Dole Soft Serve.
4. Visit the Polynesian Cultural Center to learn about Polynesian culture and the Pacific Islands.
5. Take a hike at Diamond Head Summit Trail for a panoramic view of Honolulu and the surrounding coastline.
|
|
1711.00331
| false
| null |
Our interpretability measurements are based on our proposed dataset SEMCAT, which was designed to be a comprehensive dataset that contains a diverse set of word categories. Yet, it is possible that the precise interpretability scores that are measured here are biased by the dataset used. In general, two main properties of the dataset can affect the results: category selection and within-category word selection. To examine the effects of these properties on interpretability evaluations, we create alternative datasets by varying both category selection and word selection for SEMCAT. Since SEMCAT is comprehensive in terms of the words it contains for the categories, these datasets are created by subsampling the categories and words included in SEMCAT. Since random sampling of words within a category may perturb the capacity of the dataset in reflecting human judgement, we subsample r% of the words that are closest to category centers within each category, where $r \in \lbrace 40,60,80,100\rbrace $ . To examine the importance of number of categories in the dataset we randomly select $m$ categories from SEMCAT where $m \in \lbrace 30,50,70,90,110\rbrace $ . We repeat the selection 10 times independently for each $m$ .
In contrast to the category coverage, the effects of within-category word coverage on interpretability scores can be more complex. Starting with few words within each category, increasing the number of words is expected to more uniformly sample from the word distribution, more accurately reflect the semantic relations within each category and thereby enhance interpretability scores. However, having categories over-abundant in words might inevitably weaken semantic correlations among them, reducing the discriminability of the categories and interpretability of the embedding. Table 3 shows that, interestingly, changing the category coverage has different effects on the interpretability scores of different types of embeddings. As category word coverage increases, interpretability scores for random embedding gradually decrease while they monotonically increase for the GloVe embedding. For semantic spaces $\mathcal {I}$ and $\mathcal {I}^*$ , interpretability scores increase as the category coverage increases up to 80 $\%$ of that of SEMCAT, then the scores decrease. This may be a result of having too comprehensive categories as argued earlier, implying that categories with coverage of around 80 $\%$ of SEMCAT are better suited for measuring interpretability. However, it should be noted that the change in the interpretability scores for different word coverages might be effected by non-ideal subsampling of category words. Although our word sampling method, based on words' distances to category centers, is expected to generate categories that are represented better compared to random sampling of category words, category representations might be suboptimal compared to human designed categories.
Yet, it is possible that the precise interpretability scores that are measured here are biased by the dataset used.
However, it should be noted that the change in the interpretability scores for different word coverages might be effected by non-ideal subsampling of category words. Although our word sampling method, based on words' distances to category centers, is expected to generate categories that are represented better compared to random sampling of category words, category representations might be suboptimal compared to human designed categories.
|
What are the weaknesses of their proposed interpretability quantification method?
|
can be biased by dataset used and may generate categories which are suboptimal compared to human designed categories
|
null | false
| null |
Write an apology letter to my wife for eating the last of the Oreo cookies and drinking the last of the milk.
|
To my lovely wife,
I am very sorry for consuming the last of our Oreos and Milk. I was not thinking of you or anyone else in the house when I decided to eat them but selfishly only thinking of my own hunger and needs. In the future, I promise to check with you and the kids prior to eating the last of any delicious treat or food.
My most sincere apologies.
From, Your loving Husband.
|
|
null | false
| null |
The Underground Railroad was a network of clandestine routes and safe houses established in the United States during the early- to the mid-19th century. It was used by enslaved African Americans primarily to escape into free states and Canada. The network was assisted by abolitionists and others sympathetic to the cause of the escapees. The enslaved persons who risked escape and those who aided them are also collectively referred to as the "Underground Railroad". Various other routes led to Mexico, where slavery had been abolished, and to islands in the Caribbean that were not part of the slave trade. An earlier escape route running south toward Florida, then a Spanish possession (except 1763–1783), existed from the late 17th century until approximately 1790. However, the network now generally known as the Underground Railroad began in the late 18th century. It ran north and grew steadily until the Emancipation Proclamation was signed by President Abraham Lincoln. One estimate suggests that by 1850, approximately 100,000 enslaved people had escaped to freedom via the network
|
What was the Underground Railroad?
|
The Underground Railroad was a secret network of routes and safe houses in the United States established in the early 19th century that led to free states, Canada, Mexico and other overseas areas. Black slaves used the Underground Railroad to escape from slavery. It is estimated that by 1850, about 100,000 slaves had escaped to freedom by the "Railroad".
|
null | false
| 334
|
Automated deriving discourse relation from free text is a challenging but im-portant problem. The shallow discourse parsing is very useful in the text summariza-tion BIBREF0 , opinion analysis BIBREF1 and natural language generation. Shallow discourse parser is the system of parsing raw text into a set of discourse relations between two adjacent or non-adjacent text spans. Discourse relation is composed of a discourse connective, two arguments of the discourse connective and the sense of the discourse connective. Discourse connective signals the explicit dis-course relation, but in non-explicit discourse relation, a discourse connective is omit-ted. Two arguments of the discourse connective, Arg1 and Arg2, which are the two adjacent or non-adjacent text spans connecting in the discourse relation. The sense of the discourse connective characterizes the nature of the discourse relations. The following discourse relation annotation is taken from the document in the PDTB. Arg1 is shown in italicized, and Arg2 is shown in bold. The discourse connective is underlined.
The connective identifier finds the connective word, “unless”. The arguments identifier locates the two arguments of “unless”. The sense classifier labels the dis-course relation. The non-explicit identifier checks all the pair of adjacent sentences. If the non-explicit identifier indentifies the pair of sentences as non-explicit relation, it will label it the relation sense. Though many research work BIBREF2 , BIBREF3 , BIBREF4 are committed to the shallow discourse parsing field, all of them are focus on the subtask of parsing only rather than the whole parsing process. Given all that, a full shallow discourse parser framework is proposed in our paper to turn the free text into discourse relations set. The parser includes connective identifier, arguments identifier, sense classifier and non-explicit identifier, which connects with each other in pipeline. In order to enhance the performance of the parser, the feature-based maximum entropy model approach is adopted in the experiment. Maximum entropy model offers a clean way to combine diverse pieces of contextual evidence in order to estimate the probability of a certain linguistic class occurring with a certain linguistic context in a simple and accessible manner. The three main contributions of the paper are:
The rest of this paper is organized as follows. Section 2 reviews related work in discourse parsing. Section 3 describes the experimental corpus–PDTB. Section 4 de-scribes the framework and the components of the parser. Section 5 presents experi-ments and evaluations. Conclusions are presented in the Section 6.
The shallow discourse parsing is very useful in the text summariza-tion (Marcu , 2000), opinion analysis (Somasundaran et al., 2009) and natural language generation.
|
In what aspects the shallow discourse parser is very useful?
|
Text summariza-tion, opinion analysis and natural language generation.
|
null | false
| null |
What is a Keystone?
|
A keystone is the top block of an arch. It is an essential part of an arch because it keeps other stones from falling. A keystone is shaped like a trapezoid, the widest part at the top, so it doesn't slide out of the arch. Humans have built remarkable buildings with heavy ceilings on arches supported by a keystone. Many architects and artisans have been known to sculpt family emblems, symbols and designs on keystones.
|
|
1910.00825
| false
| null |
To demonstrate SPNet's effectiveness, we compare it with two state-of-the-art methods, Pointer-Generator BIBREF5 and Transformer BIBREF6. Pointer-Generator is the state-of-the-art method in abstractive document summarization. In inference, we use length penalty and coverage penalty mentioned in BIBREF36. The hyperparameters in the original implementation BIBREF5 were used. Transformer uses attention mechanisms to replace recurrence for sequence transduction. Transformer generalizes well to many sequence-to-sequence problems, so we adapt it to our task, following the implementation in the official OpenNMT-py documentation.
To demonstrate SPNet's effectiveness, we compare it with two state-of-the-art methods, Pointer-Generator BIBREF5 and Transformer BIBREF6.
|
What are previous state-of-the-art document summarization methods used?
|
The answers are shown as follows:
* Pointer-Generator
* Transformer
|
null | false
| null |
What is Passover?
|
Passover is a major Jewish holiday that celebrates the Biblical story of the Israelites' escape from slavery in Egypt.
|
|
null | false
| null |
What state makes the best BBQ?
|
There is no one answer, but I believe Texas for beef and North Carolina for Pork.
|
|
null | false
| null |
Tell me if these are body parts or places: nose, beach, supermarket, feet, eyes, hand, mall, school, restaurant, heart
|
The body parts are: nose, feet, eyes, hand, heart
The places are: beach, supermarket, mall, school, restaurant
|
|
null | false
| null |
What is Inflation?
|
Inflation is the rate of increase in the prices of goods and services over a specified period of time. A high inflation value would represent greater increase in the prices of goods and commodities leading to lower value of the currency over-time.
|
|
null | false
| null |
Why do the Dutch wear Orange?
|
The Dutch wear orange NOT because their flag is orange (the Dutch flag is actually red, white and blue).
However, the first Kind of the Netherlands was William of Orange (Willem van Oranje), who come from the House of Orange-Naussau; one of the most influential royal house in Europe in the 16th and 17th century. Originally this house started in the Principality of Orange (France).
This principality was named after a previous Roman-Celtic Settlement (Arausio) - but it was also know for being on the route where oranges came into the country.
The family of the Prince of Orange officially adopted the name and the colour orange in 1570. After the Eighty Years' War (1568-1648) against the French, the Dutch became independent first as a republic. In 1815, the Netherlands became a monarchy under the House of Orange-Naussau.
|
|
null | false
| null |
Tell me who are the top 5 leading scorers of all time in the NBA
|
The all time leading scorers in the NBA are Lebron James, Kareem Abdul-Jabbar, Karl Malone, Kobe Bryant and Michael Jordan
|
|
null | false
| null |
What are some different types of desserts?
|
Some different types of desserts are: cupcakes, cookies, cake, chocolate fondue, pudding, and lemon tart.
|
|
null | false
| null |
What to do if you drop your iPhone into toilet?
|
I'm a bit embarrassed to admit this, but I once dropped my new iPhone into the toilet by accident. It's a humbling experience that can happen to anyone. Fortunately, the water was clean and I was able to quickly retrieve the phone. However, the screen was dark, so I used a towel to remove the water and a dryer to dry it out. I then placed the phone in a container of rice and left it overnight. The next day, when I tried turning it on, it worked! I was relieved that I didn't have to buy a new one. The lesson I learned is to be more careful and not put my iPhone near the toilet.
|
|
null | false
| null |
What US state has the most coastline?
|
Alaska is the state with the largest coastline at approximately 66,000 miles.
|
|
null | false
| null |
The Masters Tournament
The tournament has a number of traditions. Since 1949, a green jacket has been awarded to the champion, who must return it to the clubhouse one year after his victory, although it remains his personal property and is stored with other champions' jackets in a specially designated cloakroom. In most instances, only a first-time and currently reigning champion may remove his jacket from the club grounds. A golfer who wins the event multiple times uses the same green jacket awarded upon his initial win (unless they need to be re-fitted with a new jacket). The Champions Dinner, inaugurated by Ben Hogan in 1952, is held on the Tuesday before each tournament, and is open only to past champions and certain board members of the Augusta National Golf Club. Beginning in 1963, legendary golfers, usually past champions, have hit an honorary tee shot on the morning of the first round to commence play. These have included Fred McLeod, Jock Hutchinson, Gene Sarazen, Sam Snead, Byron Nelson, Arnold Palmer, Jack Nicklaus, Gary Player, Lee Elder, and Tom Watson. Since 1960, a semi-social contest at the par-3 course has been played on Wednesday, the day before the first round.
|
Based on the reference text, what does the champion of the Masters Tournament receive?
|
The champion of the Masters Tournament receives a green jacket. However, it must be returned to the clubhouse the following year to be stored in a designated cloakroom.
|
null | false
| null |
What are some popular social media applications?
|
Popular social media applications are Facebook, Instagram, TikTok, Twitter, Reddit, and Youtube
|
|
null | false
| null |
What is Bart Simpson's best friend named?
|
Millhouse
|
|
null | false
| null |
What is a cashless option exercise?
|
A cashless option exercise is a method of exercising a stock option without using cash to pay the exercise price. Instead of cash, a portion of the shares are used to pay the exercise price.
There are two methods of a cashless exercise. The first method is a net settlement, where the company withholds a sufficient number of shares with a fair market value equal to the aggregate exercise price, and such withheld shares constitute payment of the exercise price. The second method is a sell-to-cover (or same-day sale), where sufficient shares equal to the aggregate exercise price are sold and the proceeds are used to pay the exercise price. Generally, the second method is available only for companies whose shares are publicly traded, as there is greater complexity to use this method for privately held companies.
|
|
null | false
| null |
Is it safe to look at the sun directly?
|
No, it is not safe to look directly at the sun as it can cause damage to the eyes
|
|
null | false
| 222
|
We study human-human negotiations in the buyer-seller bargaining scenario, which has been a key research area in the literature BIBREF0. In this section, we first describe our problem setup and key terminologies by discussing the dataset used. Later, we formalize our problem definition.
Dataset: For our explorations, we use the Craigslist Bargaining dataset (CB) introduced by BIBREF4. Instead of focusing on the previously studied game environments BIBREF5, BIBREF6, the dataset considers a more realistic setup: negotiating the price of products listed on Craigslist. The dataset consists of 6682 dialogues between a buyer and a seller who converse in natural language to negotiate the price of a given product (sample in Table TABREF1). In total, 1402 product ad postings were scraped from Craigslist, belonging to six categories: phones, bikes, housing, furniture, car and electronics. Each ad posting contains details such as Product Title, Category Type and a Listing Price. Moreover, a secret target price is also pre-decided for the buyer. The final price after the agreement is called the Agreed Price, which we aim to predict.
Defining the problem: Say we are provided with a product scenario $S$, a tuple: (Category, Title, Listing Price, Target Price). Define the interactions between a buyer and seller using a sequence of $n$ events $E_n:<e_{1}, e_{2}, ..., e_{n}>$, where $e_{i}$ occurs before $e_{j}$ iff $i<j$. Event $e_{i}$ is also a tuple: (Initiator, Type, Data). Initiator is either the Buyer or Seller, Type can be one of (message, offer, accept, reject or quit) and Data consists of either the corresponding natural language dialogue, offer price or can be empty. Nearly $80\%$ of events in CB dataset are of type `message', each consisting a textual message as Data. An offer is usually made and accepted at the end of each negotiation. Since the offers directly contain the agreed price (which we want to predict), we only consider `message' events in our models. Given the scenario $S$ and first $n$ events $E_n$, our problem is then to learn the function $f_{n}$: $A = f_{n}(S, E_n)$ where $A$ refers to the final agreed price between the two negotiating parties.
Each posting contains the title, category, a target for the seller (called listing price), and a secret target for the buyer (called target price). The final price after the agreement is called the agreed price, which we aim to predict.
|
What does each posting contain?
|
Each posting contains the title, category, a target for the seller (called listing price), and a secret target for the buyer (called target price). The final price after the agreement is called the agreed price, which they aim to predict.
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.