`, long text divided into sections using `
`/`
`. 3. Lists use `
- `, quotes use `
`, code uses `
`, important sentences use ``.
4. Preserve all original information, no additions, deletions, or modifications.
5. No document-level tags like ``, ``, ``.
2. Analyze the Source Text:
"Dependency parsing resource development for Arabic has largely focused on Modern Standard Arabic, leaving dialectal Arabic comparatively under-resourced. We introduce MADARTB, the first large-scale parallel dependency treebank spanning five city dialects: Beirut, Cairo, Doha, Rabat and Tunis, and extending an existing parallel MSA version. In total MADARTB consists of 5,000 sentences annotated using the CATiB dependency scheme. We analyze syntactic similarity across all city dialects and MSA. Although global construction profiles exhibit high similarity across varieties, local syntactic structures frequently diverge, revealing substantial structural variation across dialects. We further establish a cross-dialect dependency parsing comparison using a supervised parser and four LLMs. The supervised parser achieves 87.5% LAS, while the best LLM setup trails at 85.7% LAS across all six varieties on the test set. We release our annotations and models publicly."
3. Drafting the HTML Structure (Iterative Process):
* *Rule Check:* Need `` or `
` for sections. Since the text is a single cohesive abstract/summary, I need to break it into logical thematic sections to justify the heading tags. Let's create three main sections:
* Section 1: Introduction & Resource Introduction (Focus on MADARTB)
* Section 2: Syntactic Analysis
* Section 3: Parsing Comparison & Results
* *Section 1 Draft:*
Dependency parsing resource development for Arabic has largely focused on Modern Standard Arabic, leaving dialectal Arabic comparatively under-resourced.
We introduce MADARTB, the first large-scale parallel dependency treebank spanning five city dialects: Beirut, Cairo, Doha, Rabat and Tunis, and extending an existing parallel MSA version.
In total MADARTB consists of 5,000 sentences annotated using the CATiB dependency scheme.
* *Section 2 Draft:*
We analyze syntactic similarity across all city dialects and MSA.
Although global construction profiles exhibit high similarity across varieties, local syntactic structures frequently diverge, revealing substantial structural variation across dialects.
* *Section 3 Draft:*
We further establish a cross-dialect dependency parsing comparison using a supervised parser and four LLMs.
The supervised parser achieves 87.5% LAS, while the best LLM setup trails at 85.7% LAS across all six varieties on the test set.
We release our annotations and models publicly.
* *Applying formatting tags:*
* Important sentences -> ``
* "the first large-scale parallel dependency treebank" -> maybe? Let's make "the first large-scale parallel dependency treebank" `` or "MADARTB" ``. Actually, let's highlight key terms like "first large-scale parallel dependency treebank", "substantial structural variation". * "87.5% LAS", "85.7% LAS" -> `` makes sense here as they are key results. * Lists -> Not strictly applicable in this short text unless I format the cities as a list. The prompt says "列表内容用
- ". I should probably use a list for the five city dialects to show I can apply this rule properly, even though it's in the middle of a sentence. Wait, rule 4 says "保留原文所有信息，不得增删删改内容". Converting "Beirut, Cairo, Doha, Rabat and Tunis" to a list changes the punctuation but preserves the info. It's safer to just keep it as is to strictly avoid altering the original text structure, OR I can make it a list. Let's stick strictly to the text to be safe, but maybe add a list for the six varieties in the last section? "across all six varieties". Let's just leave it as text to strictly follow "不得增删改内容" (no additions/deletions/modifications). I will only use `` for emphasis on crucial facts. * Wait, is there any quote? No. * Code? No. * Lists? No lists in the original, so I shouldn't invent them, but wait, the prompt says "列表内容用
- ". If there are no lists, I don't need to force one. I will just use `
` and `
/
`. * *Refining the Headings:* Let's use `
` for the main themes.
Resource Introduction
Dependency parsing resource development for Arabic has largely focused on Modern Standard Arabic, leaving dialectal Arabic comparatively under-resourced.
We introduce MADARTB, the first large-scale parallel dependency treebank spanning five city dialects: Beirut, Cairo, Doha, Rabat and Tunis, and extending an existing parallel MSA version.
In total MADARTB consists of 5,000 sentences annotated using the CATiB dependency scheme.
Syntactic Analysis
We analyze syntactic similarity across all city dialects and MSA.
Although global construction profiles exhibit high similarity across varieties, local syntactic structures frequently diverge, revealing substantial structural variation across dialects.
Parsing Comparison and Results
We further establish a cross-dialect dependency parsing comparison using a supervised parser and four LLMs.
The supervised parser achieves 87.5% LAS, while the best LLM setup trails at 85.7% LAS across all six varieties on the test set.
We release our annotations and models publicly.
4. Review against constraints: * Only output HTML snippet? Yes. * No Markdown fences? Yes. * No explanations? Yes. * Paragraphs in `
`? Yes. * Long text divided by `
`/`
`? Yes, divided into 3 `
` sections. * Lists `
- `? Not applicable as there are no explicit list items in the source. * Quotes `
`? Not applicable. * Code `
`? Not applicable.
* Important sentences ``? Used for MADARTB, substantial structural variation, 87.5% LAS, 85.7% LAS.
* Preserve all info, no add/delete/modify? Yes.
* No `//`? Yes.
5. Final Polish:
Make sure the HTML is clean and well-formed. Ensure no extra spaces or newlines that might be interpreted as markdown. Just pure HTML string.
```html
Resource Introduction
Dependency parsing resource development for Arabic has largely focused on Modern Standard Arabic, leaving dialectal Arabic comparatively under-resourced.
We introduce MADARTB, the first large-scale parallel dependency treebank spanning five city dialects: Beirut, Cairo, Doha, Rabat and Tunis, and extending an existing parallel MSA version.
In total MADARTB consists of 5,000 sentences annotated using the CATiB dependency scheme.
Syntactic Analysis
We analyze syntactic similarity across all city dialects and MSA.
Although global construction profiles exhibit high similarity across varieties, local syntactic structures frequently diverge, revealing substantial structural variation across dialects.
Parsing Comparison and Results
We further establish a cross-dialect dependency parsing comparison using a supervised parser and four LLMs.
The supervised parser achieves 87.5% LAS, while the best LLM setup trails at
`, code uses ``, important sentences use ``. 4. Preserve all original information, no additions, deletions, or modifications. 5. No document-level tags like ``, ``, ``. 2. Analyze the Source Text: "Dependency parsing resource development for Arabic has largely focused on Modern Standard Arabic, leaving dialectal Arabic comparatively under-resourced. We introduce MADARTB, the first large-scale parallel dependency treebank spanning five city dialects: Beirut, Cairo, Doha, Rabat and Tunis, and extending an existing parallel MSA version. In total MADARTB consists of 5,000 sentences annotated using the CATiB dependency scheme. We analyze syntactic similarity across all city dialects and MSA. Although global construction profiles exhibit high similarity across varieties, local syntactic structures frequently diverge, revealing substantial structural variation across dialects. We further establish a cross-dialect dependency parsing comparison using a supervised parser and four LLMs. The supervised parser achieves 87.5% LAS, while the best LLM setup trails at 85.7% LAS across all six varieties on the test set. We release our annotations and models publicly." 3. Drafting the HTML Structure (Iterative Process): * *Rule Check:* Need `` or `
` for sections. Since the text is a single cohesive abstract/summary, I need to break it into logical thematic sections to justify the heading tags. Let's create three main sections: * Section 1: Introduction & Resource Introduction (Focus on MADARTB) * Section 2: Syntactic Analysis * Section 3: Parsing Comparison & Results * *Section 1 Draft:* Dependency parsing resource development for Arabic has largely focused on Modern Standard Arabic, leaving dialectal Arabic comparatively under-resourced. We introduce MADARTB, the first large-scale parallel dependency treebank spanning five city dialects: Beirut, Cairo, Doha, Rabat and Tunis, and extending an existing parallel MSA version. In total MADARTB consists of 5,000 sentences annotated using the CATiB dependency scheme. * *Section 2 Draft:* We analyze syntactic similarity across all city dialects and MSA. Although global construction profiles exhibit high similarity across varieties, local syntactic structures frequently diverge, revealing substantial structural variation across dialects. * *Section 3 Draft:* We further establish a cross-dialect dependency parsing comparison using a supervised parser and four LLMs. The supervised parser achieves 87.5% LAS, while the best LLM setup trails at 85.7% LAS across all six varieties on the test set. We release our annotations and models publicly. * *Applying formatting tags:* * Important sentences -> `` * "the first large-scale parallel dependency treebank" -> maybe? Let's make "the first large-scale parallel dependency treebank" `` or "MADARTB" ``. Actually, let's highlight key terms like "first large-scale parallel dependency treebank", "substantial structural variation". * "87.5% LAS", "85.7% LAS" -> `` makes sense here as they are key results. * Lists -> Not strictly applicable in this short text unless I format the cities as a list. The prompt says "列表内容用
- ". I should probably use a list for the five city dialects to show I can apply this rule properly, even though it's in the middle of a sentence. Wait, rule 4 says "保留原文所有信息，不得增删删改内容". Converting "Beirut, Cairo, Doha, Rabat and Tunis" to a list changes the punctuation but preserves the info. It's safer to just keep it as is to strictly avoid altering the original text structure, OR I can make it a list. Let's stick strictly to the text to be safe, but maybe add a list for the six varieties in the last section? "across all six varieties". Let's just leave it as text to strictly follow "不得增删改内容" (no additions/deletions/modifications). I will only use `` for emphasis on crucial facts. * Wait, is there any quote? No. * Code? No. * Lists? No lists in the original, so I shouldn't invent them, but wait, the prompt says "列表内容用
- ". If there are no lists, I don't need to force one. I will just use `
` and `
/
`. * *Refining the Headings:* Let's use `
` for the main themes.
Resource Introduction
Dependency parsing resource development for Arabic has largely focused on Modern Standard Arabic, leaving dialectal Arabic comparatively under-resourced.
We introduce MADARTB, the first large-scale parallel dependency treebank spanning five city dialects: Beirut, Cairo, Doha, Rabat and Tunis, and extending an existing parallel MSA version.
In total MADARTB consists of 5,000 sentences annotated using the CATiB dependency scheme.
Syntactic Analysis
We analyze syntactic similarity across all city dialects and MSA.
Although global construction profiles exhibit high similarity across varieties, local syntactic structures frequently diverge, revealing substantial structural variation across dialects.
Parsing Comparison and Results
We further establish a cross-dialect dependency parsing comparison using a supervised parser and four LLMs.
The supervised parser achieves 87.5% LAS, while the best LLM setup trails at 85.7% LAS across all six varieties on the test set.
We release our annotations and models publicly.
4. Review against constraints: * Only output HTML snippet? Yes. * No Markdown fences? Yes. * No explanations? Yes. * Paragraphs in ``? Yes. * Long text divided by `
`/`
`? Yes, divided into 3 `
` sections. * Lists `
- `? Not applicable as there are no explicit list items in the source. * Quotes `
`? Not applicable. * Code ``? Not applicable. * Important sentences ``? Used for MADARTB, substantial structural variation, 87.5% LAS, 85.7% LAS. * Preserve all info, no add/delete/modify? Yes. * No `//`? Yes. 5. Final Polish: Make sure the HTML is clean and well-formed. Ensure no extra spaces or newlines that might be interpreted as markdown. Just pure HTML string. ```htmlResource Introduction
Dependency parsing resource development for Arabic has largely focused on Modern Standard Arabic, leaving dialectal Arabic comparatively under-resourced.
We introduce MADARTB, the first large-scale parallel dependency treebank spanning five city dialects: Beirut, Cairo, Doha, Rabat and Tunis, and extending an existing parallel MSA version.
In total MADARTB consists of 5,000 sentences annotated using the CATiB dependency scheme.
Syntactic Analysis
We analyze syntactic similarity across all city dialects and MSA.
Although global construction profiles exhibit high similarity across varieties, local syntactic structures frequently diverge, revealing substantial structural variation across dialects.
Parsing Comparison and Results
We further establish a cross-dialect dependency parsing comparison using a supervised parser and four LLMs.
The supervised parser achieves 87.5% LAS, while the best LLM setup trails at