jiazhengli's picture
Update README.md
106b953 verified
metadata
model-index:
  - name: robinlee99/long-t5-tglobal-large-AERA
    results: []
language:
  - en
base_model: google/long-t5-tglobal-large
license: apache-2.0
widget:
  - text: >-
      [Question]: 

      A group of students wrote the following procedure for their investigation.

      Procedure:

      1.Determine the mass of four different samples.

      2.Pour vinegar in each of four separate, but identical, containers.

      3.Place a sample of one material into one container and label. Repeat with
      remaining samples, placing a single

      sample into a single container.

      4.After 24 hours, remove the samples from the containers and rinse each
      sample with distilled water.

      5.Allow the samples to sit and dry for 30 minutes.

      6.Determine the mass of each sample.

      The students's data are recorded in the table below.

      A table contains four columns: Sample, Starting Mass (g), Ending Mass (g),
      Difference in Mass (g).

      The sample for the first row is Marble, with 9.8 Starting Mass, 9.4 Ending
      Mass and -0.4 for Difference in

      Mass.

      The sample for the second row is Limestone, with 10.4 Starting Mass, 9.1
      Ending Mass and -1.3 for Difference

      in Mass.

      The sample for the third row is Wood, with 11.2 Starting Mass, 11.2 Ending
      Mass and 0.0 for Difference in Mass.

      The sample for last row is Plastic, with 7.2 Starting Mass, 7.1 Ending
      Mass and -0.1 for Difference in Mass.

      After reading the group's procedure, describe what additional information
      you would need in order to replicate

      the experiment.

      Make sure to include at least three pieces of information.


      [Key Elements]:

      Needed Information:

      You need to know how much vinegar was used in each container.

      You need to know what type of vinegar was used in each container.

      You need to know what materials to test.

      You need to know what size/surface area of materials should be used.

      You need to know how long each sample was rinsed in distilled water.

      You need to know what drying method to use.

      You need to know what size/type of container to use.

      Other acceptable responses.


      [Marking Rubric]:

      3 points: The response describes three additional pieces of information
      that would be needed to accurately

      replicate the experiment;

      2 points: The response describes two additional pieces of information that
      would be needed to accurately

      replicate the experiment;

      1 point: The response describes one additional piece of information that
      would be needed to accurately

      replicate the experiment;

      0 point: The response describes little or no accurate or relevant
      information from the acid rain investigation.


      [Student Answer]: The procedures I think they should have included inorder
      for me to replicate the experiment would be how different samples did they
      used for each? What tool did they use to determine the mass.

      [Score and Rationale]:
    example_title: AES Eassy Set 1
  - text: >-
      [Question]: 

      A student performed the following investigation to test four different
      polymer plastics for stretchability.

      Procedure:

      1. Take a sample of one type of plastic, and measure its length.

      2. Tape the top edge of the plastic sample to a table so that it is
      hanging freely down the side of the table.

      3. Attach a clamp to the bottom edge of the plastic sample.

      4. Add weights to the clamp and allow them to hang for five minutes.

      5. Remove the weights and clamp, and measure the length of the plastic
      types.

      6. Repeat the procedure exactly for the remaining three plastic samples.

      7. Perform a second trial (T2) exactly like the first trial (T1).

      The student recorded the following data from the investigation.

      The table shows the amount of stretch (in millimeters) for four different
      types of plastic, labeled as A, B,

      C, and D, when subjected to two different stretching forces, labeled as T1
      and T2.

      For plastic type A, it stretched 10mm under T1 and 12mm under T2.

      For plastic type B, it stretched 22mm under T1 and 23mm under T2.

      For plastic type C, it stretched 14mm under T1 and 13mm under T2.

      Lastly, for plastic type D, it stretched 20mm under both T1 and T2.

      a. Draw a conclusion based on the student’s data.

      b. Describe two ways the student could have improved the experimental
      design and/or validity of the results.


      [Key Elements]:

      Conclusions:

      Plastic sample B has more stretchability than the other polymer plastics.

      Plastic sample A has the least amount of stretchability compared to the
      other polymer plastics.

      Not all polymer plastics have the same stretchability.

      Different polymer plastics have different stretchability (and are
      therefore suited for different applications).

      A reasonable conclusion cannot be drawn due to procedural errors.

      Other reasonable conclusions 


      Experimental Design Improvements:

      Provide the before and after measurements for length (Did the samples all
      start out the same size?).

      Make sure the samples are all of the same thickness.

      Variations in thickness could have caused variations in stretchability.

      Perform additional trials.

      Some of the samples have similar stretchability (A and C, B and D).

      Two trials may not be enough to conclusively state that one is more
      stretchable than the other.

      Indicate how many weights were added to the clamps (Was it the same number
      for each sample?).

      Other acceptable responses


      [Marking Rubric]:

      3 points: The response draws a valid conclusion supported by the student’s
      data and describes two ways the

      student could have improved the experimental design and/or the validity of
      the results;

      2 points: The response draws a valid conclusion supported by the student’s
      data and describes one way the

      student could have improved the experimental design and/or the validity of
      the results. -or- The response

      describes two ways the student could have improved the experimental design
      and/or the validity of the results

      but fails to draw or incorrectly draws a conclusion from the student’s
      data;

      1 point: The response draws a valid conclusion supported by the student’s
      data but fails to describe, or

      incorrectly describes, how the student could have improved the
      experimental design and/or the validity of the

      results. -or- The response describes one way the student could have
      improved the experimental design and/or

      the validity of the results but fails to draw or incorrectly draws a
      conclusion from the student's data.;

      0 points: The response provides little or no correct information from the
      polymer investigation.


      [Student Answer]: Based on the students data I can conclude that plastic
      type B has the largest amount stretched. Two ways that the student could
      have improved the experimental decison is by allowing the weights to hang
      longer to see if the plastic would stretch more. Another way he could have
      improved his experiment is by doing a third trials to really be sure that
      his results are correct.

      [Score and Rationale]:
    example_title: AES Eassy Set 2
  - text: >-
      [Question]: 

      Starting with mRNA leaving the nucleus, list and describe four major steps
      involved in protein synthesis.


      [Key Elements]:

      mRNA exits nucleus via nuclear pore.

      mRNA travels through the cytoplasm to the ribosome or enters the rough
      endoplasmic reticulum.

      mRNA bases are read in triplets called codons (by rRNA).

      tRNA carrying the complementary (U=A, C+G) anticodon recognizes the
      complementary codon of the mRNA.

      The corresponding amino acids on the other end of the tRNA are bonded to
      adjacent tRNA’s amino acids.

      A new corresponding amino acid is added to the tRNA.

      Amino acids are linked together to make a protein beginning with a START
      codon in the P site (initiation).

      Amino acids continue to be linked until a STOP codon is read on the mRNA
      in the A site (elongation and

      termination).


      [Marking Rubric]:

      3 points: Four key elements;

      2 points: Three key elements;

      1 point: One or two key elements;

      0 points: Other.


      [Student Answer]: 1. tRNA tranfers it outside the nucleus to be matched
      with a ribosome2. The ribosome latches on to an amino acid.3. The codons
      match up4. It creates protein.

      [Score and Rationale]:
    example_title: AES Eassy Set 5
  - text: >-
      [Question]: 

      List and describe three processes used by cells to control the movement of
      substances across the cell membrane.


      [Key elements]:

      Selective permeability is used by the cell membrane to allow certain
      substances to move across.

      Passive transport occurs when substances move from an area of higher
      concentration to an area of lower

      concentration.

      Osmosis is the diffusion of water across the cell membrane.

      Facilitated diffusion occurs when the membrane controls the pathway for a
      particle to enter or leave a cell.

      Active transport occurs when a cell uses energy to move a substance across
      the cell membrane, and/or a

      substance moves from an area of low to high concentration, or against the
      concentration gradient.

      Pumps are used to move charged particles like sodium and potassium ions
      through membranes using energy and

      carrier proteins.

      Membrane-assisted transport occurs when the membrane of the vesicle fuses
      with the cell membrane forcing large

      molecules out of the cell as in exocytosis.

      Membrane-assisted transport occurs when molecules are engulfed by the cell
      membrane as in endocytosis.

      Membrane-assisted transport occurs when vesicles are formed around large
      molecules as in phagocytosis.

      Membrane-assisted transport occurs when vesicles are formed around liquid
      droplets as in pinocytosis.

      Protein channels or channel proteins allow for the movement of specific
      molecules or substances into or out of

      the cell.


      [Marking Rubric]:

      3 points: Three key elements;

      2 points: Two key elements;

      1 point: One key element;

      0 points: Other.


      [Student Answer]: osmosis- when water is diffused across a membrane from
      more concentrated to less concentrated.diffusion- materials that are
      crossed over from the membraneactive transport

      [Score and Rationale]:
    example_title: AES Eassy Set 6

NEW!! A newer version of this project is avaliable at here.

Model Card for long-t5-tglobal-large-AERA

This repository provides a fine-tuned version of long-t5-tglobal-large, using our proposed AERA framework presented in the paper: Distilling ChatGPT for Explainable Automated Student Answer Assessment.

Citation Information

@inproceedings{li-etal-2023-distilling,
    title = "Distilling {C}hat{GPT} for Explainable Automated Student Answer Assessment",
    author = "Li, Jiazheng  and
      Gui, Lin  and
      Zhou, Yuxiang  and
      West, David  and
      Aloisi, Cesare  and
      He, Yulan",
    editor = "Bouamor, Houda  and
      Pino, Juan  and
      Bali, Kalika",
    booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
    month = dec,
    year = "2023",
    address = "Singapore",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2023.findings-emnlp.399",
    doi = "10.18653/v1/2023.findings-emnlp.399",
    pages = "6007--6026",
    abstract = "Providing explainable and faithful feedback is crucial for automated student answer assessment. In this paper, we introduce a novel framework that explores using ChatGPT, a cutting-edge large language model, for the concurrent tasks of student answer scoring and rationale generation. We identify the appropriate instructions by prompting ChatGPT with different templates to collect the rationales, where inconsistent rationales are refined to align with marking standards. The refined ChatGPT outputs enable us to fine-tune a smaller language model that simultaneously assesses student answers and provides rationales. Extensive experiments on the benchmark dataset show that the proposed method improves the overall QWK score by 11{\%} compared to ChatGPT. Furthermore, our thorough analysis and human evaluation demonstrate that the rationales generated by our proposed method are comparable to those of ChatGPT. Our approach provides a viable solution to achieve explainable automated assessment in education",
}