Professional Documents
Culture Documents
Chapter 1
INTRODUCTION
1.1 BACKGROUND
Since the emergence of the earliest writing systems, textual instructions have always
been among the most ubiquitous means of transferring procedural knowledge.
Nowadays, free instructions are available on the Web for numerous tasks of human
life, be they easy every-day hacks, recipes, or professional-level instructions on using or
repairing tech- nical devices of the highest complexity. For the latter, in- structions
might be written in a complex and technical style, and they, in general, often use situated
language (Malmaud et al., 2014). The instructional text is regularly divided into
multiple steps, which should be performed in a specified se- quence. Consequently, there
could be a sequence of objects that the person interacts with during the process.
Extract- ing such practical information from text presents a unique challenge in the
domain of information extraction (IE).
The type and structure of the extracted information also de- pend on the characteristics
of the task and the intended use of the information. In this work, we focus on IE from a
highly complex type of instructions: repair manuals. More specifically, we focus on
extracting the pieces of informa- tion that a collaborative repair assistant would benefit
most from, in order to support a human in a repair task.
A fruitful collaboration among agents depends on multiple prerequisites, one of
them being sufficient shared knowl- edge among the agents. Many researchers have
stressed the importance of shared knowledge and representation struc- ture in
cooperative situations (Grice, 1975; Salas et al., 1995). In our case, we are mainly
interested in how this in- formation builds up and is spread across the different steps as
well as across the different sentences in a single step. More precisely, we focus on the
sequence and identity of the required tools and disassembled parts during the repair
process. This information can help to estimate the state of the task and workstation
environment, serving as the “sit- uational context” of task-oriented dialogue, as it is
defined in Deutsch (1974).
When a task-oriented collaboration involves objects in the shared workspace of
the agents, an essential communica- tive function is resolving the partner’s references
to ob- jects in the environment. Wilkes-Gibbs and Clark (1992) indicated that when
people collaborate on referring expres- sions, they issue full noun phrases initially,
after which they begin to shorten the phrases. Thus, in the process of re- pair, one
could say “Give me the T8 Torx screwdriver”, and then “Give me the screwdriver”,
or just “next one”. In such situations, the task knowledge helps the partner to
anticipate the required objects and disambiguate the refer- ences, so that the speaker
can be less explicit in referring expressions (Whitney et al., 2016). Generally, when the
agents share a common source of knowledge, such as the task instructions, it can be
expected that their cooperation becomes smoother and more intuitive.
1.2 MOTIVATION
Today’s world is an era of accelerated technological progress characterized by new
innovations whose rapid application and diffusion typically cause an abrupt change in
society. The evolution of computer, mobile, networks, the devices that run on them and
their everyday services occur at an amazing rate. It is unthinkable to consider our lives
without mobile phones. Mobile phones have been one of the most successful
technologies ever invented and adopted in the ever-developing world. Apart from
making everyday life easy, mobile phones, computers and internet are most common
weapons used by criminals to commit heinous crimes (McSweeney, 2020). These
weapons are commonly collected evidences in cybercrimes that are examined by
investigators. Sometimes criminals intentionally damage their mobile phones and
computers to destroy the evidence. Therefore, it’s becoming more challenging for an
investigator/ examiner to extract data from the evidences (Dongan & Akbal, 2017).
1.4 SOLUTION
In our approach we used we introduce a semi-structured dataset of repair manuals. The
dataset is annotated in a large category of devices, with information that we consider
most valuable for an automated repair assistant, including the required tools and the
disassembled parts at each step of the repair progress. We then propose methods that
can serve as baselines for this IE task: an unsupervised method based on a bags-of-
n-grams similarity for extracting the needed tools in each repair step, and a deep-
learning-based sequence labeling model for extracting the identity of disassembled parts.
These baseline methods are integrated into a semi-automatic web-based annotator
application that is also available along with the dataset.
Chapter 2
LITERATURE SURVEY
We gathered the manuals from iFixit along with their meta- data in JSON-like objects,
where each document has self- explaining tags regarding the Title, Category, a list of
all required tools provided by the instructor, which we refer to as the Toolbox, a list of
Steps, and a derived list of hierar- chical categories for the device (Ancestors). In most
cases, the instructors show step-by-step procedures for opening the device and
removing or repairing a broken component. Since reassembly is usually the reverse of
disassembly, it is not included in the guides of iFixit. Each step has an at- tribute Lines,
which contains the text description of the step and, if available, the attribute Image with
the link(s) to the provided image(s). The fact that 98.7 % of the steps have one or more
images also makes this data suitable for multi- modal studies. Figure 1 shows an
instance of the dataset, including the extra tags we added to the data that contain the
annotated information in this work.
DATA STATISTICS
In total, 31,601 repair manuals were collected from the iFixit API in 15 basic
categories, see Figure 2. There is a high variation in the number of steps
(average=9.68, me- dian=7.00, variance = 109.95), depending on the category of the
device and the difficulty level of the task. However, there is less variation in the
number of tools pre-specified in the toolboxes, where the average number is 2.42 tools
per manual with median is 2.00 and variance is 3.98.
This selection was due to it having the highest average number of steps and tools per
manual, 24.5 and 4.3, respectively. With an approximate estimation based on the
specified level of dif- ficulty and the required repair time in iFixit, these devices
appeared to be among the most complicated devices for re- pair. One should note that
in iFixit, the instructors often copy the text description of some steps from other
manuals in order to create a new guide, and hence, there are steps with identical text
descriptions. In fact, there are 4,350 steps with unique text descriptions in this category.
There- fore, we manually annotated the steps with unique text de- scriptions and then
applied the same annotation to the steps with identical text. Table 1 shows the statistics
of annotated data.
Chapter 3
SYSTEM ANALYSIS
Figure 2: Parts of the Damaged Mobile Phone after Disassembling the Device
The mobile phone (Fig.3) (Wiens & 19 other contributors)Samsung, Model: SM-
N975F/DS, had two layered motherboard, like a silicon sandwich (Fig.4) (Wiens & 19
other contributors). This mobile device was then repaired using the forensic repair tool
kit that restored the device in its working condition and data extraction became possible.
Case opening:
The case was received from a messenger in sealed condition. The parcel was
appropriately marked. The parcel contained one mobile phone of Samsung, Model:
N975F/DS (in Broken and Damaged condition). The mobile phone was appropriately
marked/ tagged for identification.
As the device was found in damaged condition, the device was physically examined with
the help of a microscope after examining through naked eyes. The device was taken
apart, into pieces by unscrewing it. Except motherboard every part of the device was
damaged i.e. the screen, back panel, screen parts, etc. were all broken. Data extraction
was possible for this mobile device by repairing it. Hence, materials for repairing were
required. New screen combo was purchased for the device and all the parts of the
exhibit were transplanted and installed with the new screen combo. The mobile phone
was kept for charging but it showed no indication for the process. Hence, a look at PCB
was taken as the theory pointed towards a possible PCB short-circuiting.
There are aspects in the feasibility study portion of the preliminary investigation:
Technical feasibility
Economic feasibility
Operational feasibility
This study is carried out to check the technical feasibility, that is, the technical
requirements of the system. To develop this system, we first worked with web
applications and then it would be technically feasible. The requirements here are very
modest because the system here supports very basic window version which is 8 and
above. So, this system is technically feasible because it needs only window with a basic
version.
HARDWARE COSTS
SOFTWARE COSTS
MONGO DB FREE
The fully computerized will reduce the man hours And overhead by simplifying by today
activities. Thus the computerized system is beneficial for anyone & so it is economical
feasible. It can be done with help of COCOMO Model known as constructive cost
model.
Ab Bb Cb Db
Calculation
KLOC=3.0
Effort per Month (E) = ab*(KLOC) ^bb
= 3.6*(3.0) ^1.20
E = 13.45(approx) person- Month
Development Time(D)=cb*(E)^db
=2.5*(13.45) ^0.32
=5.74 month
No. of Team Members = E/D
= 13.45/5.74
= 2.89
=~3 person’s approx
For personal cost we did the following calculations:
Generally the computer engineer has Rs.16000 salary per month, therefore
1 month = Rs.15000
Per day he works only 8 hours
Therefore 30*8=240 hours
He works 240 hours per month
Therefore,
15000/240=Rs 66.6 for 1 hour
We are three project partners. We work for 5 hr a week i.e. 5 weeks a month
The aspect of the study is to check the level of acceptance of the system by the
user. This includes the process of training the user to use the system efficiently. This
system will not threaten the user instead it is friendly in its operation.
i. August Month
1.Feasibility Study
2. Analysis of
Modules
3. Inclusion of
Modules
4. Compilers
Survey
1.Rough Sketch of
GUI
2. Researching of
Communication
Process
3. Study of servlets
and TCP/IP Comm.
4. Preparation of
basic module
. Work
Week 1 Week 2
Tasks
1.Creation of UML
diagrams of the
design
2. Preparation of
presentati
Chapter 4
SYSTEM REQUIREMENT SPECIFICATION
4. Mongo DB- It is an local Database where the data is stored in the form of tables.
It mainly analyzes the student record and gives criteria to acquire the performance.
1. All the contents in the project are implemented using Graphical User Interface (GUI).
2. Every conceptual part of the projects is reflected using the Data mining.
3. System gets the input and delivers extracted knowledge that describes students'
behavior.
4.3.1.2 Software Interfaces
1. Using Java Script we have created popupfrontend tool design and React JS for
coding.
2. Mongo DB database is used for back end to store the exam result and retrieve in the
graph view.
Data Mining is the concept of getting the needed information from the large datasets. It
is one of the steps involved in Knowledge Discovery in Database. The other steps in
Knowledge Discovery in Database such as Selection, Transformation, Interpretation,
Preprocessing.
Chapter 5
SYSTEM DESIGN
Processing of
Data
5.1.1 MODULES
1. Signing up
Creating a User account is the first step for anyone to access the website. New
users should choose the signup option in the launcher page and provide the necessary
details and register into website.
2. Login Screen
Once the user is registered they can use the username and password to log into the
website for further Mobile Assesories will be display for Solving Problem issue of
Mobile.
Once the valid user enters into website, it ask for name to be identified in the list
of available users .
4. Database
The Database would contain the list of available Problems of Mobile. The
details stored as, username and password for an authentication. After solving the issue of
Mobile details stored in the database are retrieved in the form of graph.
5. Processing
In the beginning of the case, the forwarding authorities were stumped by the ingenious
planning of the smugglers who had left no clues behind that could directly connect them
to the crime. The suspects had already managed to destroy any implicating evidence of
the crime and had even tried to destroy their mobile devices. However, the innovative
and industrious efforts of the scientific officers of the case enabled a complete
restoration of the mobile device. The forwarding authority’s invaluable support by
providing the password and an intricate work in repairing the device by the scientific
officer led to the successful solution of the case.
6. Exit
To close the app, the user should select the exit option from the main
layout.
Mobile Handset
Mobile Configuration
Customer
ASSESSMENT
Start
No Authenti Yes
cation
Solution to Repair
Mobile Handset
End
Stop
Use case describe the interaction between one or more actors and the system
itself, represented as a sequence of simple steps that take part in a sequence of activities
in a dialog with the system to achieve goal.
Mobile Problem
Handset Configure.
Repair Bill
ADMIN
REGISTRATION
REGISTER_USER()
ALLOCATED_ID()
DISPLAY DATA
GENERATION
CUSTOMER
MOB_INFO
CUSTOMER_ID
MOBILE_NAME()
MODEL()
FEATURES()
REPAIR SOLUTION
UNIQUE_ID
READ DATA()
CHECK DATA()
DISPLAY ID()
REGISTER
ASSIGN UNIQUE ID
SEND INFO
SEND DETAILS
GENERATE RESULT
CUSTOMER
REGISTER
ADMIN
MOBILE PROBLEM
WAIT FOR
SOLUTION
Report
Close
Chapter 6
IMPLEMENTATION
Step 1: Start.
Step 2: Register to the website.
Step 3: click on login button.
Step 4: New Customer can register in register page.
Step 5: Once the Customer registered, provide the user name and password in Login
page.
Step 1: Start
Step 2: Click on start button.
Step 3: Customer submit problem issue of Mobile Handset & wait for Solution.
Step 4: Finally get Solution.
Chapter 7
SCHEDULE OF WORK
1. Gender recognition
using facial image.
2. TV show popularity
analysis using data
mining.
3. “An Annotation
Tools Base Line
Chapter 8
8.1 CONCLUSION
We have presented the first steps for information extrac- tion from repair manuals and
towards an intelligent repair assistant. We introduced a dataset of repair manuals with
semi-structured information. The dataset is annotated in the category of Mac Laptops,
with the objects that a user might interact with in the course of the repair task. The
dataset is well-suited for studying information extraction from long and technical
passages, where the arguments and actions might be elided, implicit and spread across
several passages. We then proposed methods that can serve as base- lines for determining
the required tools and the disassem- bled parts in each repair step. The methods, along
with mul- tiple NLP techniques, are integrated into a semi-automatic annotation web-
based tool, which is freely available in ad- dition to the dataset.
In this paper, we only extract the disassembled parts us- ing the word-level
annotation of steps, i.e. when the iden- tity of the disassembled part can be extracted by
labeling the corresponding words in the text. For future work the method needs to be
enhanced for learning the step-level in- formation, where extracting the information
goes beyond the word labeling in descriptions and often requires extra knowledge.
objects is a dis- tinctive property of repair manuals. When we detach a component from a
device using a tool, eventually, typically after multiple intermediate steps, we have to
reattach that component, most likely with the same tool. The specific or- der of
components in the devices, which is partially shared among devices in a category, makes
it feasible to attempt predicting the next step in the repair task, e.g., by using pre- dictive
models for learning statistical scripts, such as Hid- den Markov Models (Orr et al., 2014)
or LSTM Neural Net- works (Pichotta and Mooney, 2016). Finally, we plan to use the
extracted information as the context for a task-oriented dialog in a human-robot
interaction scenario and we en- visage that this dataset and associated baseline
methods are similarly beneficial for work on other smart assistance systems and their
constituent NLP and task-understanding components.
BIBLIOGRAPHY